Part 1 of this series examined the September 2025 protest events of Nepal in the days leading up to and during the violence itself. This section shifts focus to ask what made those days possible. Why were tens of thousands of young Nepalis willing to risk confrontation with security forces? What had shifted in Nepal’s political and economic landscape to make this level of mobilization feasible? For atrocity prevention practitioners, this case study on Nepal serves to demonstrate why structural risk assessment must extend beyond monitoring immediate conflict indicators, and why understanding how these vulnerabilities interact and compound over time matters significantly for early warning frameworks. 

Nepal’s Labor Market

Nepal’s labor force is predominantly informal and agriculture-based. Nearly three-fifths of the economically active population is engaged in agriculture, even though agriculture contributes only around 25% of GDP.  Post-COVID, the government and donors have advocated for modernizing agriculture (e.g. agro-processing, commercialization) to improve productivity and youth appeal. 

Nepal’s industrial sector (including manufacturing, construction, utilities, mining) employs roughly 15% of workers. Construction is a major employer, accounting for about 14% of employment (nearly 1 million jobs in 2018).

The services sector in Nepal employs about 20–22% of the workforce, but generated over 60% of GDP in the fiscal year 2024/25. Notably, tourism and hospitality is a significant sub-sector: just before COVID, tourism directly employed about 371,000 people (11.5% of total employment), making it the fourth-largest industry by employment.

Nepal’s digital economy is an emerging sector that has shown promise, particularly for urban youth. Currently, information and communication services employ less than 1% of workers (around 60,000 people as of 2018). This sector grew during and after the pandemic as businesses and education went online. The business process outsourcing (BPO) industry – such as call centers, IT support, and software services – is expanding rapidly. 

Nepal has a low-cost, English-speaking workforce, and even though the BPO/IT sector is small relative to the labor force, it is regarded as a “fast-emerging sector” with increasing demand for tech skills. The government and development partners identify digital services and ICT as a high-growth area for job creation. In fact, the World Bank’s 2025 strategy for Nepal highlights digital jobs (e.g. IT services, digital finance) as a priority for absorbing educated youth.  

The Unemployment Trap

Nepal has one of the youngest populations in South Asia. Approximately 56% of its 30 million citizens are under the age of 30. Yet, the country’s labor market has been structurally incapable of absorbing its own young workforce, leaving an entire generation trapped between educational aspiration and economic exclusion.

Youth unemployment reached 20.82% in 2024, more than double the overall unemployment rate of 10.71%. Over the past three decades, youth unemployment has consistently hovered around twice the national average. A large share of youth work in family farms or informal jobs that are not officially counted. The NEET rate (youth neither in employment, education or training) is among the highest in South Asia, especially for young women. 

On the bright side, the government’s investments in roads and energy have translated into thousands of construction jobs, though many are temporary. The manufacturing sector, while small, has seen pockets of growth (e.g. in food processing and construction materials) that need young factory workers and technicians. Additionally, telecommunications and retail (such as mobile phone services, shopping centers) grew as consumers turned to e-commerce and delivery services during the pandemic, which opened new roles for youth in delivery, sales, and customer service. 

Remittance Dependency

Over the past two decades, Nepal’s reliance on remittances has deepened significantly. In 2024, remittance inflows reached approximately 26% of GDP, positioning Nepal among the nations with the highest remittance-to-GDP ratios globally. Totaling over $11 billion annually, these funds serve as the country’s primary source of foreign exchange and provide essential financial support to millions of households. The World Bank’s post- Gen Z revolution Development Update estimates that without the inflow of remittances, an additional 2.6 million people would fall below the poverty line. 

About 14% of the population work abroad, primarily in Gulf states (Qatar, Saudi Arabia, the United Arab Emirates), Malaysia, India, South Korea, and Japan. At the national level, remittances support the balance of payments and shore up foreign exchange reserves, preventing a currency crisis. 

Foreign Direct Investments (FDI) & Political Instability

The persistent imbalance in Nepal’s external trade remains one of the most salient indicators of its economic vulnerability. Despite exports experiencing remarkable growth of 58 percent year-on-year in the first five months of fiscal year 2025/26, they represented merely 13 percent of total trade.

In the last fiscal year, pledged FDI amounted to approximately $489 million, while net inflows reached only about $90 million, indicating that less than one-fifth of committed investment materialised. Analysts and former officials at the Central Bank of Nepal identify procedural delays, licensing and pre-approval requirements, weak regulatory coordination, corruption, and political instability as key deterrents to investment execution. 

Since Nepal abolished its monarchy and became a federal democratic republic in 2008, power has frequently alternated among a small group of prominent leaders from the three largest parties: K.P. Sharma Oli (CPN-UML), Sher Bahadur Deuba (Nepali Congress), and Pushpa Kamal Dahal (‘Prachanda’). Critics and, increasingly, ordinary citizens see this as a political class that treats governance as a game of musical chairs, where power is rotated among the same faces with little regard for the aspirations or welfare of the population. 

Rural Neglect and Disaster Vulnerability

Nepal’s underinvestment in rural infrastructure and disaster preparedness has left millions of citizens dangerously vulnerable to natural hazards. In 2024 and 2025 alone, Nepal experienced a series of catastrophic disasters. On July 12, 2024, two buses with 65 passengers were swept into the Trishuli River by a landslide on the Madan Ashrit Highway. In September of the same year, severe flooding and landslides across more than 50 districts including Kavrepalanchok, Sindhuli, and Lalitpur resulted in 224 deaths and extensive destruction. More recently, in October of 2025, heavy rainfall triggered landslides and flooding in Ilam district, killing at least 44 people and leaving several missing.

Despite these recurring disasters, the government has made efforts to strengthen disaster management. With support from international donors and NGOs, Nepal enacted the Disaster Risk Reduction and Management Act of 2017, which established a multi-tier institutional structure to coordinate disaster response across federal, provincial, and local governments. Following the devastating 2015 earthquake, Nepal updated building codes and strengthened their enforcement in affected districts. However, the continued scale of casualties and destruction suggests that implementation gaps, inadequate funding, and insufficient rural infrastructure investment continue to undermine these disaster preparedness efforts.

Nepal’s Digital Economy and Social Media Influence

Social media did not cause Nepal’s 2025 crisis, but it fundamentally reshaped how grievances were articulated, amplified, and acted upon.

Over the past five years, Nepal has experienced a significant expansion in digital connectivity. By early 2025, internet penetration had reached 56 percent, with 16.5 million individuals using the internet. Social media adoption has also expanded rapidly, with 14 million active social media user identities recorded in January 2025 equivalent to 48 percent of the total population. These trends point to a markedly more connected and digitally mediated social environment than in previous years. Content creation, digital entrepreneurship, freelancing, and online businesses have become viable pathways to income, particularly for urban youth with digital literacy and reliable internet access. Digital advertising now commands 34% market share, surpassing television (28%), print media (18%), radio (12%), and outdoor advertising (8%). The dominance of digital advertising is driven by increased internet and smartphone use, the effectiveness of social media campaigns, and the growing influence of the creator economy.

However, this digital growth is unevenly distributed. Internet access varies dramatically by economic status: only 9.5% of those living in poverty have internet access. Similarly, cable TV access is 9% for poor households versus 32% for non-poor households. Geographic disparities are equally pronounced: while 79% of households in Kathmandu Valley have internet access, this drops to 43% in other urban areas and just 17% in rural regions.

Social media did not cause Nepal’s 2025 crisis, but it fundamentally reshaped how grievances were articulated, amplified, and acted upon. The convergence of widespread smartphone adoption, algorithmic amplification, and a politically frustrated youth population created the conditions for a digitally mediated uprising that ultimately forced a transition in the country’s political leadership. The 2025 protests were largely led by youth in urban areas who had reliable internet access, digital literacy, and the ability to navigate platforms like TikTok, Instagram, Reddit, and Discord. Rural populations, by contrast, were less digitally connected and less able to participate in online coordination efforts. This does not mean rural grievances were any less severe, but it does demonstrate that the movement’s digital infrastructure was concentrated in cities, particularly Kathmandu. 

Nepal’s Gen Z protests cannot be understood as a sudden phenomenon driven by social media alone, nor as a purely economic revolt. Rather, it emerged from the cumulative interaction of long-standing vulnerabilities that compounded over time and narrowed peaceful channels for political expression. For early warning and atrocity prevention frameworks, Nepal’s case study highlights the need to move beyond event-driven indicators and toward deeper assessments of how stressors interact to transform latent discontent into mass mobilization.

Part 3 of this series turns from diagnosis to the use of early warning/early action technologies. It explores the emerging ecosystem of geospatial early-warning tools, conflict forecasting models and mapping platforms that can be used by governments, civil society organizations, and international actors in atrocity prevention. Drawing on the applied work of the Atrocity Prevention Lab, it asks how spatial data, risk models, and community-generated information can be translated into timely early action rather than reaction.

---

Reflections on Geospatial Artificial Intelligence for Atrocity Prevention

GeoAI is increasingly positioned as a transformative tool for humanitarian early warning and early action. GeoAI is the integration of geospatial data and artificial intelligence to analyze, model, and predict spatial patterns and processes. Geospatial data is used to train models to recognize patterns, classes, and features, enabling more effective spatial analysis and interpretation. From satellite imagery to incident reporting systems, spatial analytics promise to identify emerging risks before violence escalates. Yet, in practice, the effectiveness of these systems is constrained not by algorithms alone, but by the nature of the data and the ethical responsibilities attached to its use.

These reflections emerge from applied work developing early warning web maps and analytical tools for atrocity prevention in South Sudan, as a graduate student trainee at the Harvard Humanitarian Initiative’s Atrocity Prevention Lab.

The Reality of Humanitarian Data

Humanitarian datasets rarely arrive in an ideal form. Incident records are often compiled under conditions of insecurity, limited access, and uneven reporting capacity. In many cases, datasets lack accompanying documentation, contain ambiguous variable encodings, or rely on legacy data entry systems.

Variables such as gender or sex may be numerically coded without definition. Categories describing killings, kidnappings, sexual violence, or torture may be embedded within victim attributes rather than contextual dynamics. These are not merely technical inconveniences; they shape what GeoAI systems can learn and how risk is ultimately represented.

Before early warning can responsibly inform early action, foundational questions about data meaning and structure must be confronted.

Modeling Risk or Modeling Visibility?

One of the central challenges in atrocity prevention analytics is distinguishing between actual risk and the capacity to report violence. Areas with stronger reporting networks may appear more volatile, while remote or marginalized regions remain underrepresented.

When GeoAI models are trained on such data without correction, they risk predicting visibility rather than vulnerability. Early warning systems may then reinforce existing blind spots rather than illuminate emerging threats.

Addressing reporting bias is not optional; it is essential for ethical and effective atrocity prevention.

The Ethics of Prediction Under Uncertainty

In conflict settings, uncertainty is unavoidable. Ground truth may be delayed, incomplete, or politically sensitive. Yet early warning systems often translate probabilistic outputs into seemingly definitive risk maps.

This raises critical ethical questions. How should uncertainty be communicated to decision-makers? What level of confidence is sufficient to justify intervention—or restraint? And how do we prevent overconfidence in models when the consequences of error can be severe?

Responsible GeoAI must make uncertainty visible rather than obscure it.

From Early Warning to Early Action

Despite advances in spatial modeling, many humanitarian systems remain stuck in early warning mode. Risk is identified, mapped, and reported—but not always acted upon.

This gap is rarely technical alone. Institutional workflows, governance structures, and decision-making thresholds often determine whether early warning leads to preventive action. Without alignment between analytical outputs and operational realities, even the most sophisticated models risk irrelevance.

Early action requires not just better models, but better integration between analytics and humanitarian decision-making.

Who Defines the Variables That Matter?

Gender, sex, weapon type, and perpetrator categories are frequently treated as standard variables. Yet their definitions vary across contexts and datasets. When these inconsistencies are absorbed into GeoAI pipelines without scrutiny, they can introduce analytical noise or reinforce harmful assumptions.

Humanitarian GeoAI must remain attentive to how variables are defined, by whom, and for what purpose. Technical efficiency should not override contextual integrity.

Accountability in Humanitarian AI

As GeoAI increasingly informs atrocity prevention efforts, accountability becomes more complex. When a model fails to identify an emerging threat or falsely signals risk, responsibility is often diffused.

Clear governance frameworks are needed to define accountability across data collection, model development, interpretation, and decision-making. Without them, humanitarian AI risks becoming authoritative without being accountable.

Centering Human Judgment

GeoAI should not replace human expertise. Field analysts, local practitioners, and affected communities provide essential contextual knowledge that no model can fully capture.

The most effective early warning systems are those that amplify human judgment rather than displace it using GeoAI as a decision-support tool, not a decision-maker.

Conclusion

GeoAI holds significant promise for atrocity prevention and humanitarian early action. Yet its success depends less on computational power than on ethical clarity, data integrity, and institutional responsibility. Early warning is not simply a technical challenge. It is a humanitarian one.

---

In November 2025, the Atrocity Prevention Lab (APL) hosted a three-part webinar series bringing together global leaders to tackle some of the toughest challenges in atrocity prevention—from shrinking resources to disappearing data and rapidly evolving technologies. Below is a summary of the key insights, lessons, and calls to action that emerged across the three conversations.

Webinar 1 – Expanding the Frontiers of Atrocity Prevention with Spatial Strategies

The first webinar in the series, hosted on November 5th, 2025, examined the application of geospatial analysis as an evidentiary tool to enforce legal mandates and operationalize atrocity prevention. Speakers emphasized the urgent need to bridge the gap between early warning and effective response by turning static maps into actionable mechanisms for legal accountability. The session moved beyond technical demonstrations to address how data can be refined to trigger decisive interventions in volatile regions like the Sahel and the Democratic Republic of the Congo.

Tammy Palacios–Senior Fellow at Middlebury Institute, Research Fellow at West Point, and CEO of Priority Sustainable Solutions LLC–opened the dialogue by focusing on the professionalization of prevention efforts. She stressed the importance of identifying specific data needs, ensuring that decision-makers receive the right information at the right time. Discussing a report on the Sahel, Tammy noted that while hundreds of local Civil Society Organizations were mapped in the region, they remained largely disconnected from major funding bodies like USAID. Her argument centered on the necessity of connecting these local actors with policymakers, suggesting that sustainable mapping solutions are essential to bridge the silos between grassroots realities and high-level strategy.

Ronald Serwanga of East Africa Law Society, International Bridges to Justice, Fountain Advocates, and Loyola University Chicago–shifted the focus to the legal application of these tools, proposing a synergistic model to integrate spatial technologies with international and domestic legal frameworks. Using the eastern DRC as a case study, he demonstrated that displacement is not random but geographically predictable, often correlating with mineral sites. Ronald argued that while robust mandates exist, such as the 2025 DRC Land Use Law, there is often a lack of objective triggers to enforce them. He posited that geospatial evidence serves as a critical evidentiary link, converting subjective community warnings into objective evidence that renders the crisis structural and actionable before UN bodies.

Addressing the operational realities of this work, Gillian Elliott–Project Coordinator at the Signal Program–discussed conflict early warning in an environment shaped by significant reductions in US support. She presented alarming data from a mixed-methods survey indicating that permanent funding cuts are threatening the survival of a majority of organizations managing conflict early warning systems in Sub-Saharan Africa. Gillian highlighted the human cost of this retrenchment, noting how years of hard-earned community trust are unraveling, citing the example of a health clinic built over eight years that collapsed when funding ceased. Her segment underscored the urgent need for solidarity and network strengthening as organizations are forced to reduce staff and services, cancel programs, and seek new donors.

The webinar concluded with a broad call for innovation amidst these resource constraints, emphasizing the imperative of maximizing impact with limited means. The discussion highlighted that in a landscape of diminishing funding, strategic networking and cost-effective tools are non-negotiable. Speakers highlighted the power of visual data to alert decision-makers, noting that it serves as a persuasive mechanism to drive advocacy. Ultimately, the session ended on a collaborative note, urging the community to engage in joint research and share resources to maintain the momentum of atrocity prevention work.

Webinar 2 – Knowledge Commons and Global Practices for Atrocity Prevention

The second webinar in the series, hosted on November 12th, 2025, highlighted critical threats to global atrocity prevention efforts posed by the loss of institutional knowledge and data. Experts across academia, fieldwork, and network mapping shared insights on how political shifts, funding changes, and staffing turnovers are eroding vital “knowledge commons” essential for early warning, risk assessment, and sustained prevention action.

The keynote speaker, Dr. Samantha Lakin–Genocide Studies Scholar and Senior Fellow at the Center for Peace, Democracy, and Development at UMass Boston–spoke on knowledge commons and global practices for atrocity prevention. She stressed the real-time disappearance of institutional memory through vanishing online repositories and departing experts, and advocated for collaborative, multi-institutional efforts to document and preserve knowledge as a form of resistance against these losses.

The next speaker, Peter Mwamachi–a leading early warning expert and experienced peace-building practitioner based in Kenya–emphasized practical challenges in Kenya, such as shifting government priorities and resource gaps. He urged embedding early warning systems within local and national frameworks to enhance resilience.

The final speaker, Nate Haken–Consultant and Senior Advisor to Fund For Peace–presented a decade-long social network mapping initiative across West Africa that visualizes the invisible infrastructure, trust networks, and collaborations fundamental to coordinated atrocity prevention. This network memory, if preserved, prevents fragmentation of prevention efforts and strengthens collective action.

While the knowledge preservation challenge is complex, the panelists agreed on clear ways forward: 

1. Non-cloud backups, 
2. Cross-sector collaboration,
3. Institutional partnerships, and
4. Building systems that outlive individual personalities. 

The webinar concluded with a call for a global, collective commitment to safeguarding knowledge commons as public goods that underpin peace and security efforts worldwide.

Webinar 3 – Innovating at the Intersection of Technology and People 

The third and final webinar in the series, hosted on November 19th, 2025, explored how tech and people can work together to strengthen conflict and atrocity prevention. Speakers stressed a simple but powerful point: tools like geospatial analysis, AI, and advanced models only work when grounded in human judgment and local expertise.

Keynote speaker Jessie Pechmann–Humanitarian GIS and Data Protection Lead at the Humanitarian OpenStreetMap Team–walked participants through wildly different estimates of buildings damaged in southern Lebanon and Gaza, depending on which dataset or method you look at. Automated building footprints and machine learning often missed or misclassified damage, while carefully edited OpenStreetMap data produced more complete and reliable baselines. Her examples from Lebanon, Gaza, and earthquake-affected Myanmar showed how local knowledge—like knowing which areas were bulldozed before an earthquake or what undamaged architecture actually looks like—is essential to interpreting satellite imagery and avoiding misleading “hard numbers.”

Dr. Roudabeh Kishi–Affiliate Faculty at the University of Denver and Founder & Director of Crisis Lens–focused on what it takes to make early warning models useful beyond the usual conflict hotspots. Drawing on her experience at ACLED, she noted that many AI models perform reasonably well where there’s lots of past violence to learn from, but they struggle in more stable contexts and with rarer forms of harm like gender-based violence. Her message: effective early warning needs new kinds of indicators—such as polarization, disinformation, and community reporting—backed by better data standards, ethical governance, interoperability, and much stronger data literacy so decision makers actually trust and use the outputs.

Food security expert Julien Jacob–of Action Against Hunger Spain–then shared a conflict and hunger platform developed with academic and private-sector partners. The tool combines environmental, economic, conflict, and food security data in some of the world’s most violence-affected countries to understand how specific types of attacks—on crops, markets, warehouses, or supply chains—translate into hunger. By restructuring open-source conflict data, scraping local media, using remote sensing, and training predictive models, the team is filling key data gaps in hard-to-reach areas. Julien emphasized that local teams and communities still play a crucial role in explaining what the data misses—like fear that keeps people from fields or markets—and in turning model insights into real-world decisions.

Across the webinar, the takeaway was clear: technology can dramatically expand what we can see, measure, and forecast, but it doesn’t replace human judgment, ethics, or local experience. Building better prevention means using AI and spatial tools responsibly, investing in data systems that can actually “talk” to each other, and keeping people—especially those closest to the risks—firmly in the loop from data collection to action.

Conclusion

Together, these three webinars underscored a shared message: atrocity prevention depends on more than powerful tools—it requires durable systems, trusted relationships, and a commitment to act before violence becomes inevitable. From spatial strategies that turn displacement patterns into legal triggers, to knowledge commons that safeguard institutional memory, to human-centered approaches to AI and early warning, the series highlighted how evidence, ethics, and local expertise must move in lockstep. The challenge ahead is to professionalize prevention work even in times of shrinking resources: connecting local actors to policymakers, preserving network memory across regions, and ensuring that early warning systems are not just technically sophisticated, but politically and socially grounded.  

We invite you to turn these insights into concrete next steps. Strengthening atrocity prevention means investing in resilient data and knowledge infrastructures, building cross-sector partnerships that outlast individual champions, and centering communities and local experts in every phase—from data collection and analysis to decision-making and advocacy. Whether you are a practitioner, policymaker, researcher, or funder, we encourage you to share these webinars with your networks, integrate their lessons into your own practice, and join or form collaborative efforts that break down silos across sectors.

Recordings of all three sessions are available to watch on demand at the link below; we hope you will revisit them, use them in trainings and discussions, and continue expanding this community of practice dedicated to preventing atrocities before they occur.

---

Part 1: Anatomy of Nepal’s Gen Z Revolution

On September 8, 2025, for the first time in Nepal’s modern political history, mass civilian deaths occurred in an anti-corruption movement that carried neither party banners nor ideological manifestos.

Instead, the movement was driven by a leaderless network of young students, coordinated through digital platforms and united by frustration over corruption, censorship, and unaccountable governance.

The state’s response on September 8 constituted one of the most severe episodes of violence against civilians in Nepal’s post-1990 democratic era. The gathering that day was widely expected to be non-violent given that it was composed mostly of teenagers and students in school uniforms and because public protest had long been an institutionalized feature of Nepali civil society. Instead, it became the deadliest single day in Nepal’s democratic history, claiming the lives of more young, unarmed civilians on a single day than during the entire 2006 pro-democracy movement.

By the end of September, at least 75 civilians had been confirmed dead and over 2,000 injured.

By the end of September, at least 75 civilians had been confirmed dead and over 2,000 injured. The majority of fatalities resulted from gunshot wounds inflicted during clashes between demonstrators and security forces; forensic and medical reports indicate that nearly all gunfire victims were struck above the waist, primarily in the head, neck, and chest, strictly violating crowd control protocols. Several deaths also occurred as a result of arson-related incidents during the rioting, with victims trapped in government buildings, shops, and residential structures that were set ablaze. Emergency responders struggled to reach affected areas as daytime curfews paralyzed much of Kathmandu, and health facilities were overwhelmed by the influx of patients. Tear gas was deployed at Civil Service Hospital, where victims were being brought for treatment, further worsening the humanitarian situation.

This three-part analysis examines the 2025 Nepal uprising through three central questions:

1. Part 1: What were the immediate triggers and key events of the movement that led to the government’s collapse?
2. Part 2: What pre-existing structural and economic fractures, along with weaknesses in digital governance made Nepal vulnerable to this crisis?
3. Part 3: How can geospatial technologies be leveraged for civilian protection and atrocity prevention in Nepal?

Origins of the Protest Movement

Public discontent in Nepal reached a boiling point on September 4, 2025, after the government banned 26 social media platforms for failing to comply with its registration requirements. The move was the final spark for a public already enraged. For months, a viral trend on social media (#NepoBaby) had brought widespread attention and commentary on the strong disparity between the lavish lifestyles of politicians and their families and the public that had long been frustrated by rampant corruption, unemployment, and human capital flight; this Gen Z movement had already spread through Indonesia, the Philippines, and Bangladesh. Following a pattern seen across Southeast Asia, this online movement finally erupted into the streets of Nepal when the government pulled the plug on their primary means of expression.

The requirements for social media registration were rooted in the Social Media Management Directive 2080 in November 2023, under the authority of the Electronic Transaction Act 2063, which required digital platforms serving Nepalese users to register locally by establishing a physical presence in the country and appointing grievance and compliance officers to monitor content and ensure legal accountability. Despite multiple warnings and diplomatic outreach, most platforms did not meet the registration deadline of 28 August 2025.

Officials framed the shutdown as a long-overdue regulatory intervention, arguing that unregistered platforms had become breeding grounds for hate speech, coordinated misinformation campaigns, online harassment, and content that could incite violence. However, critics said the ban went far beyond routine regulation and was aimed at silencing online dissent.

As anger moved from social media to the streets, several locations across the Kathmandu Valley quickly became the main arenas of confrontation. The protests began at Maitighar Mandala, a major roundabout and a site of public demonstrations in central Kathmandu, before moving east toward New Baneshwor and the Federal Parliament Building via Bijuli Bazaar. As the unrest widened, curfews were imposed around Shital Niwas (the President’s Office) and Singha Durbar (central administrative complex of the Government of Nepal), while violence also spread to Balkot in Bhaktapur, the area surrounding former Prime Minister Oli’s private residence.

Timeline of Nepal’s Gen Z Movement:

September 8, 2025: The Protest and Initial Crackdown

• 9:00 a.m. – Thousands of young protesters began gathering at Maitighar Mandala in central Kathmandu to rally. By 9 a.m., the Maitighar intersection was filled with school and college students (many in uniform) and young professionals carrying banners like “Youth Against Corruption.” Some students were also seen waving anime One Piece flags.
• 10:00 – 11:00 a.m. – As crowds grew, protesters marched east toward New Baneshwor, singing patriotic songs along the way. By 11:38 a.m., demonstrators breached the first police barricades, prompting security forces to deploy water cannons. Tear gas followed at 11:55 a.m., and by 12:08 p.m., around twenty individuals had climbed the main gate of the Federal Parliament.
• 12:30 p.m. – The Kathmandu District Administration Office imposed a curfew in New Baneshwor and surrounding areas, though much of the crowd could not hear the announcement. At 12:37 p.m., police opened fire for the first time, targeting protesters who had climbed onto the terrace of a small building near the Parliament gate.
• 12:45 – 2:00 p.m. – By 12:50 p.m., a dense crowd had formed near the west-facing Parliament gate. At 12:57 p.m., protesters climbed a police checkpost inside the Parliament compound, and at 1:20 p.m., a Naya Patrika journalist was shot while reporting. Injured protesters were transported to the Civil Service Hospital as several vehicles burned outside and at least one ambulance was damaged.
• 2:30 – 3:30 p.m.– Around 3:00 p.m., police entered the Civil Service Hospital premises and fired tear gas inside the compound. By 3:30 p.m., preliminary reports confirmed fourteen fatalities.
• 5:00 – 6:00 p.m. – By early evening, officials reported nineteen protesters killed and more than 100 injured. Prime Minister Oli convened an emergency cabinet meeting at 6:00 p.m.
• 8:00 – 11:30 p.m. – Home Minister Ramesh Lal Lekhak resigned at 8:00 p.m. Late that night, at 11:25 p.m., Communications Minister Prithvi Subba Gurung announced the lifting of the social-media ban that had sparked the protests. Curfews were expanded to Shital Niwas, the Vice-President’s residence, and the Singha Durbar complex. The army was placed on standby as unrest spread to other cities overnight.

September 9, 2025: Continued Unrest and PM Oli’s Resignation

• Morning – Despite curfews, fresh protests erupted across the country. At 10:07 a.m., Agriculture Minister Ramnath Adhikari resigned. Former Chief Justice Sushila Karki visited Civil Service Hospital at 10:36 a.m., followed by Health Minister Pradeep Poudel’s resignation at 10:45 a.m. By 11:42 a.m., former Prime Minister Sher Bahadur Deuba’s residence had been set on fire, and around noon some members of Rastriya Swatantra Party (RSP) announced their collective resignation. Offices of the Nepali Congress and CPN-UML were also torched.
• 12:30 – 1:00 p.m. – At 12:54 p.m., former Prime Minister Oli’s private residence in Balkot was set ablaze, followed six minutes later by the burning of Gagan Thapa’s home. Additional arson was reported at the Department of Transport Management and the Hilton Hotel in Kathmandu.
• 2:00 – 3:30 p.m. – Prime Minister Oli formally resigned from office. At 2:30 p.m., security officials publicly stated they had no authorization to open fire even if major government buildings were attacked. Fires then spread to the Supreme Court at 3:00 p.m., the Maoist Party headquarters at Paris Danda at 3:08 p.m., and later to the District Court and Special Court.
• 4:00 – 10:00 p.m. – Singha Durbar was reported to be burning heavily, and the President’s residence at Shital Niwas had also come under attack. Over 3,000 inmates broke out of Central Jail Sundhara between 8:00 p.m. and 10:00 p.m.

This detailed chronology of the September 8-9, 2025 uprising is synthesized from extensive ground reporting and an interactive map published by Onlinekhabar.

Constitutional Crisis and Interim Governance

In the aftermath of ex-PM Oli’s resignation on September 9, the Nepal Army invited representatives of the Gen Z movement for dialogue. During the meeting, Chief of Army Staff Ashok Raj Sigdel mentioned businessman Durga Prasai and the Rastriya Swatantra Party (RSP) as stakeholders in the movement, which prompted several Gen Z representatives, including Rakshya Bam, to walk out. According to Republica’s reporting, Bam stated outside Army Headquarters that the Army Chief had instructed them to meet the President and participate in negotiations together with Durga Prasai and the RSP faction. She said the delegates believed that being grouped with them would undermine the credibility of their movement, so they refused the proposal and left the meeting.

Shortly after, on a Discord server called “Youth Against Corruption” with over 145,000 members, more than 10,000 users met virtually to debate potential candidates through hours of discussion, multiple polls, and the use of sub-channels for fact-checking. Five names were shortlisted for the final voting: Harka Sampang, a social activist and mayor of Dharan; Mahabir Pun, a popular social entrepreneur and activist running the National Innovation Centre; Sagar Dhakal, an independent youth leader; advocate Rastra Bimochan Timalsina (known as Random Nepali on YouTube); and former Chief Justice Sushila Karki. An open Discord poll drew 7,713 votes, with Karki receiving 3,833 votes (50 percent).

The selection process involved three days of hectic negotiations, with sharp differences among different groups over appointing Karki as prime minister. While President Ramchandra Paudel was adamant about staying within constitutional bounds, Gen Z youths were frustrated over what they saw as unnecessary delays, with differences centering on whether and when to dissolve Parliament. On September 12, President Ram Chandra Poudel appointed Karki as interim Prime Minister under Article 61(4) of the Constitution of Nepal, making her the country’s first female head of government. Upon her recommendation, Poudel dissolved the 275-seat parliament and set elections for March 5, 2026, approximately two years earlier than planned.

The immediate appointment of former Chief Justice Sushila Karki to the Prime Minister’s office triggered a constitutional dispute, as the decision raised a prima facie contradiction with Article 132(2) of the Constitution of Nepal. This provision explicitly states that any individual who has once held the office of Chief Justice or a Judge of the Supreme Court “shall not be eligible for appointment to any government office, except as otherwise provided in this Constitution.” This clause is a foundational provision designed to maintain the strict separation of powers and protect judicial independence by preventing the executive branch from offering future positions to members of the judiciary.

In defense of the appointment, legal experts including prominent constitutional lawyer Bipin Adhikari advocated for a pragmatic interpretation of the constitution in light of the extraordinary circumstances. Adhikari did not dispute the textual conflict with Article 132 but argued that the nation’s stability and the imperative to resolve a deadly political vacuum justified the move. His defense rested on the doctrine of necessity, suggesting that an action technically violating a specific clause could be permissible to save the constitutional order from total collapse.

Human Rights Violations

The ACLED data (Figure 1) show that public demonstrations have been a defining feature of Nepal’s political landscape for over a decade, peaking during moments of transition such as the 2015 constitutional crisis and the 2020–2021 parliamentary disputes. Still, the violence against civilians seen on 8 September 2025 was unusual.

Police logs revealed that 13,182 rounds of ammunition had been fired over the two days- 2,642 live bullets, 1,884 rubber rounds, and 6,279 tear-gas shells.

The Model Protocol for Law Enforcement Officials, endorsed by UN special rapporteurs, states that police should facilitate and protect assemblies, use negotiation and communication to de-escalate tensions, and give clear warnings and safe avenues for dispersal before resorting to force. It also calls for civilian‑led oversight and special care when children are present. These standards were not followed on September 8th’s protest.

Nepal’s Ministry of Health and Population confirmed the death of a 12-year-old student, and at least nineteen students were reported killed that day alone. A forensic report prepared by Tribhuvan University Teaching Hospital confirmed that nearly all gunshot victims were struck above the waist, primarily in the head, neck, and chest. Police logs revealed that 13,182 rounds of ammunition had been fired over the two days—2,642 live bullets, 1,884 rubber rounds, and 6,279 tear-gas shells—in violation of Nepal’s own crowd-control regulations, which requires warnings before fire and mandates that any live rounds be aimed below the knee. These actions also contravened the UN Basic Principles on the Use of Force and Firearms, which limit the use of lethal force to instances of imminent threat to life and require proportionality and precaution.

Simultaneously, mass prison breaks across 28 facilities resulted in 14,043 inmates and juvenile detainees fleeing custody; despite intensive operations, over 5,000 individuals remain at large. Over 1,200 police firearms—including INSAS rifles, SLRs, and service pistols—along with roughly 100,000 rounds of ammunition, were looted from police barracks and government offices during the unrest. As of early November 2025, more than 500 of these weapons and a significant portion of the ammunition remained unaccounted for.

Nepal’s security agencies such as Nepal Police, the Armed Police Force, the Nepali Army, and the National Investigation Department have identified escaped inmates and stolen weapons as the country’s most serious security threats ahead of the March 5 parliamentary elections. Joint patrols by the army, police, and armed police have been deployed nationwide in an attempt to restore public confidence, stabilize local conditions, and deter armed groups or criminal networks seeking to exploit the situation. However, more than 400 police stations and offices were burnt or destroyed during the unrest, forcing many units to operate from damaged or temporary structures. The Armed Police Force reported damage to 62 of its facilities, and Nepal Police estimated losses of personnel property alone at over Rs 220 million (US $1.5 million).

These incidents have drawn increased scrutiny from international human-rights organizations. The World Organization Against Torture (OMCT) and CIVICUS urged Nepal to lift curfews, withdraw the military and conduct prompt, impartial investigations. Human Rights Watch emphasized that the use of firearms in crowd control is unjustified. Given the security crisis and eroded public trust, the imperative for transparent and impartial investigations becomes urgent and non-negotiable.

Part 2 of this series will present a macro-level analysis of the structural factors that increased Nepal’s vulnerability to conflict in 2025, focusing on rapidly changing digital-information systems and long-standing development constraints.

---

Red Dot Foundation and the Safecity Platform

Atrocities rarely begin with mass killings. One of the earliest warning signs is gender-based violence (GBV). Sexual assault, domestic abuse, forced marriage, and harassment are not only grave human rights violations in their own right, but they also reveal the deeper structural inequalities that can spiral into broader persecutions and mass atrocities, if left unchecked. GBV exists on a continuum of violence and recognizing its patterns is critical to prevention. Data, when used responsibly, can help communities identify risks and intervene early. Red Dot Foundation (RDF) is one such organization, combining spatial analysis with an understanding of local social norms to address GBV and strengthen community resilience.

RDF developed Safecity, a crowdsourced platform and movement that collects, analyzes, and visualizes anonymous reports of sexual and gender-based violence (SGBV) to make public spaces safer for women and marginalized communities. Built at the intersection of gender rights, technology, and urban planning, Safecity’s methodology turns invisible experiences into actionable insights for individuals, communities, and institutions.

Safecity in Faridabad: From Data to Action

In October 2023, in Faridabad, India, Safecity piloted the Safer Faridabad project that demonstrated how anonymous community reporting, paired with institutional collaboration, can transform both public safety and survivor trust. The intervention began with safety audits and surveys conducted in identified hotspots, where over 4,000 narratives of unsafety and SGBV were collected. These narratives revealed everyday patterns of harassment such as stalking, catcalling, and indecent exposure, as well as less visible risks tied to alcohol and drug use in public spaces.

Geotagging and analyzing these reports allowed Safecity to visualize trends and create responsive programming to address emerging issues in the community. For example, while chain snatching and robbery were the most commonly reported crimes, categories like “other” began to surface new safety risks, such as harassment linked to intoxication and drug use. These insights were compiled into data dashboards and reports shared with the Faridabad police, turning invisible community experiences into actionable insights.

Community Engagement: Turning Data into Collective Action

Safecity’s approach goes beyond data dashboards. After identifying hotspots through safety audits and crowdsourced reports, Safecity facilitators engaged directly with residents, schools, and civic groups.

Workshops and awareness sessions gave people space to reimagine public safety and to co-create solutions. Safecity’s “talking boxes” model (inspired by Polycom for Girls) also ensured inclusivity in contexts where digital access is limited. In communities without reliable technology, residents could drop paper reports into secure boxes, keeping reporting accessible and anonymous.

Safecity’s approach understands that technology alone cannot shift norms or restore trust. Any solutions to address GBV must be shaped by the lived experiences of those affected. Furthermore, data analysis must be combined with dialogue, empathy and local ownership to see sustained improvements in communities.

Collaboration with Police

One of the more novel contributions of this project is the meaningful engagement and partnership with local law enforcement to address GBV. The shift to a dialogue with the police–from passive dashboards to interactive feedback sessions–led to them reviewing reports not just as individual cases but as aggregated patterns that required systemic solutions.

When confronted with data showing low reporting rates, where only around 20% of survivors filed police reports, officers participated in empathy and survivor-centered training. Through role-play, they practiced small but meaningful shifts in interaction: offering water, asking questions with compassion, and avoiding victim-blaming. Over a three-month period, intent to report to police increased by 30%, highlighting how community engagement combined with responsive police training could directly influence survivor trust.

In response to trends emerging under reports categorized as “other,” police deployed CCTV cameras in areas where drug and alcohol use correlated with higher harassment incidents, demonstrating how community data could drive rapid preventive action.

Traditional policing often responds after an incident has happened. The project differed a lot because we focused on prevention, presence and listening to people.

Quote from an interview between Signal Team and Safecity Team on July 23, 2025

Lessons Learned

The Faridabad project underscored several key insights:

• Movement patterns matter. Violence correlates with how people navigate public spaces; peaks in harassment reports often coincided with typical transit times.
• Legal clarity drives reporting. Clear-cut crimes like robbery are more frequently reported than intimate violations such as indecent exposure, where stigma and legal ambiguity remain barriers.
• Trust is foundational. Survivors may not seek linear justice; many want safety and dignity over a formal legal process. Empathetic policing helps close this gap.
• Community input is essential. Safecity’s strength lies not in its technology alone but in combining crowd-mapped data with community voices and institutional accountability

Why This Matters for Atrocity Prevention

Gender-based violence does not exist in isolation and often foreshadows broader communal violence. Safecity’s work in Faridabad illustrates how addressing GBV through data-driven, community-informed interventions can strengthen atrocity-prevention toolkits. By turning individual experiences into systemic insights, Safecity helps both communities and institutions respond before violence escalates. In centering survivor voices and using data to build trust, Safecity offers a replicable model for bridging the gap between communities and institutions in fragile contexts.

This article represents the independent analysis of the Harvard Humanitarian Initiative’s Atrocity Prevention Lab and Red Dot Foundation and does not reflect the official views of the Haryana Police or the Government of Haryana.

We would like to thank SafeCity members ElsaMarie DSilva, Camila Gomide, Sonali Alves, Soumyaa Hariharan, and Tania Echaporia for their input.

Shwetha Srinivasan is a current Research Assistant at the Harvard Humanitarian Initiative’s Signal Program on Human Security and Technology.

---

Jakarta, June 2025

Pondok Kacang Timur (PKT) is a village in Indonesia with a fairly active recitation group. There is a religious leader who is well known for giving religious lectures in the neighborhood. Unfortunately, every time she preaches and circulates her sermons through the WhatsApp Group, she always spreads negative propaganda against the government, blaming the government for being infidel, liberal, etc. This has made some people uncomfortable and disturbed, as the propaganda has created social divisions between those who are pro and those who are opposed. Following the hate speech, she invited people to join a ‘so-called’ better governance system under a khilafah—which is a political system led by a Caliph and guided by Islamic law.

In the past two decades, conflict in Indonesia has transformed from large-scale episodes of communal and separatist conflicts to more small-scale acts of violence, including violent extremism.¹ Between 2021 and 2022, Jakarta’s Center for Strategic and International Studies’ Collective Violence Early Warning (CVEW) dataset recorded 2,335 collective violence incidents which led to over 662 deaths, 2,918 injuries, and 724 instances of damaged or destroyed infrastructure.² If left unaddressed, even small-scale conflict events could escalate into larger, more lethal mass atrocity crimes.³ Indonesia faces numerous risk factors that could potentially lead to violent conflicts. These factors include the rise of intolerant acts against minority groups, the presence of extremist groups, and historical human rights abuses. Preventing the escalation of these risks as early as possible is critical to minimizing irreversible damage, including loss of life.

Violent conflict does not suddenly manifest within communities.

The government employs a monitoring system, which is generally slow to respond and thus delays action. The current system largely relies on open-source information regarding violent incidents from the news, which means that responses are implemented after a conflict event has already occurred. The existing system hardly contributes to preventive measures at early stages, and communities are minimally engaged by decision-makers, despite being at the forefront of conflict. Furthermore, there is no local mechanism through which communities can participate in early warning activities to inform timely atrocity prevention and early response. Violent conflict does not suddenly manifest within communities. Coordinated planning, the spreading of hate speech, and the build-up to violent events take place over time. The elongated period during which this happens allows early warning signs to emerge within the communities in which these activities are taking place. National interventions cannot always be implemented at a very local level, yet it is critical to capture EW signs at their inception to prevent harm to civilians.

While the current EW/EA system is centrally managed by the national and local governments, Empatiku is promoting another solution—an EW/EA application which is managed by local communities. The role of the local community as frontline responders is essential because they can more easily identify unrest and abnormalities and can thus respond in a timely manner before conflict escalates. This community-based system should shorten the decision-making process and speed up responses to early warning signs before they escalate. The civil society organization behind this application, Empatiku, has been working in Indonesia since 2015 to address the rise of violent extremism and ensure that civil society is involved in bringing about social change.⁴

This is something only a local EW/EA system is able to accomplish.

To support community decision-making processes, Empatiku is currently developing a digital EW/EA application. The utilization of information and communication technologies will support local communities to rapidly and easily document and monitor reported cases and obtain geospatial analytic reports that facilitate early response, coordination, collaboration, and communication. The role of technology in EW/EA is critical for mitigating and intervening in instances of violent conflict. The system allows community members to report early warning signs, including the identification of potentially vulnerable populations affected by conflict. Community members can input potential cases into the application. Alternatively, community leaders can collect the reports.

The Resilient Team will then manage and analyze the data to subsequently increase the effectiveness and efficiency of early action (EA) at the community level. The community-based EW/EA approach shifts agency to local communities, prioritizes community needs and accountability, and further enhances the understanding of the risks and EW indicators of violent conflict. This, therefore, improves community resiliency and conflict prevention outcomes. Furthermore, it will shorten the duration between reports of EW signs and EA, thus minimizing or even negating the harmful effects of violent conflict. This is something only a local EW/EA system is able to accomplish.

Back at the PKT village, the Resilient Team responded when the early case was reported. They discussed the case after having verified it and decided to personally approach the religious leader and communicate the residents’ concerns over the hate speech that has polarized the community. Even though she seemed to be listening and promised to change, unfortunately, she continued to spread the hate speech propaganda. The Resilient Team then referred the case to the chairman of the Mosque Welfare Council. The case was discussed, and the Mosque Welfare Council made the decision to temporarily prohibit the religious leader from preaching until she was willing to stop spreading hate speech. After several months of inactivity, the woman has finally changed.

Mira Kusumarini is the Founder and Director of EMPATIKU. She is also a member of the APL Steering Committee.

---

References:

1. Patrick Barron, Sana Jaffrey, and Ashutosh Varshney, “How Large Conflicts Subside: Evidence from Indonesia,” Indonesian Development Paper, No. 18 (2014), 1
2. Dr. Lina Alexandra and Alif Satria, “Policy Brief – Indonesia’s Collective Violence Trends in 2022: Less Common, More Lethal, Centre for Strategic and International Studies (CICS), February 2023, 2
3. Gerry van Klinken, Communal Violence and Democratization in Indonesia: Small Town Wars, (New York: Taylor & Francis, 2007), 31; Nancy Lee Peluso, “Violence, Decentralization, and Resource Access in Indonesia,” Peace Review: A Journal of Social Justice, No. 19 (2007), 24-25; Patrick Barron, Sana Jaffrey, and Ashutosh Varshney, “When Large Conflicts Subside: The Ebbs and Flows of Violence in Post-Suharto Indonesia,” Journal of East Asian Studies, No. 16 (2016), 191- 192; Yuhki Tajima, The Institutional Origins of Communal Violence: Indonesia’s Transition from Authoritarian Rule, (New York: Cambridge University Press, 2014), 8-10; United Nations Office on Genocide Prevention and the Responsibility to Protect, Framework of Analysis for Atrocity Crimes: A Tool for Prevention, (New York: United Nations, 2014), 18
4. Learn more at: https://empatiku.or.id/en/beranda/

To state the obvious: no crisis exists in a vacuum. Whether it begins with a pandemic, a natural disaster, or political unrest, the ripple effects inevitably reach across social, economic, political, and security systems. A health crisis, for example, is never just a health crisis. And the same goes for any emergency. The hard truth is that preparedness means being ready to act on multiple fronts at once—everywhere, with limited resources, and diminishing influence as events escalate. This is where advanced analytics and generative AI come in.

It is much harder to predict the cascading impacts of the crisis after it occurs.

Predictive models are of little use if all they do is tell you the location where a particular event will occur. That’s the easy part, whether through regressions, time series, cluster analysis, event history, geospatial methods, Bayesian analysis, AutoRegressive Integrated Moving Averages (ARIMA), or any combination of these and other approaches. It is easy to determine which countries are most susceptible to fires, floods, earthquakes, or terrorist attacks. It is harder to predict the timing and severity. It is much harder to predict the cascading impacts of the crisis after it occurs.

But policy makers need to prioritize response options somehow, often without a solid empirical basis to guide them. As a result they tend to 1) default to precedent (“fighting the last war”), 2) defer to influence-based decision making (deference to the “loudest voice in the room”), or 3) simply react (playing “whack-a-mole”), reluctant to allocate scarce resources on things that are not already escalating.

To address this challenge, Fund for Peace has partnered with SAS to develop a Crisis Sensitivity Simulator, to help decision-makers and responders prioritize the use of their limited resources to maximum effect. The simulator does not predict the onset of a crisis, but it takes nineteen years of historical Fragile States Index data for 179 countries, to determine how a given crisis is likely to unfold in a particular country, if and when the shock occurs.

The tool employes a three-step process as follows:

Step 1. The first step of the process uses a 4-parameter beta distribution to model country-specific factor behavior under normal conditions. It then applies shocks to individual factors, runs multiple simulations, and visualizes the results through box plots, comparing baseline and shocked distributions to assess a country’s sensitivity to different types of crises.

Step 2. Then the tool analyzes the correlations between different types of shocks over time in each individual country in order to help subject matter experts infer how a given crisis may cascade across different sectors.

Step 3. Finally, the tool integrates these analytics into a customized prompt generator, which leverages a Large Language Model (LLM) to produce detailed, localized crisis scenarios. This process accounts for both current conditions and historical trends, generating nuanced narratives that illustrate how a crisis may unfold across different communities. By incorporating data-driven insights with natural language generation, the simulator helps decision-makers anticipate cascading impacts, assess vulnerabilities, and craft more targeted response strategies. This tool showcases how responsible and trustworthy AI can optimize decision-making in complex humanitarian scenarios.

Importantly, this tool is not a replacement for local knowledge.

Importantly, this tool is not a replacement for local knowledge. Rather it is used for brainstorming and “red teaming”: helping teams explore a range of potential outcomes and cascading effects and facilitating alternative analysis by challenging assumptions and considering different perspectives. However, when we tested it in Port Harcourt, Nigeria, for a crisis scenario planning exercise on catastrophic flooding with humanitarian, peace, and development practitioners, the participants found the analytics and the AI enhancements useful for validating and amplifying their own foundation of expertise that they brought to the question for prioritizing strategic preparedness and response.

In May, we will be presenting the tool at SAS Innovate 2025, which is an annual conference focusing on advancements in data analytics and artificial intelligence that is attended by tech luminaries sharing their latest breakthrough ideas and projects. However, the real potential of the tool is not found supporting the intelligentsia, but that it makes advanced analytics accessible to local experts and practitioners in the field so that they can prioritize and target their preventative efforts and limited resources for maximum effect, in a world where the intensity and frequency of global shocks are increasing. This is all the more important at a time when guardrails and stabilizing mechanisms like international norms, multilateral cooperation, and evidence-based policymaking are being tossed aside by the most powerful country in the world.

Nate Haken is an expert in early warning and conflict sensitivity and a member of the APL Steering Committee.

---

The escalation of the Lebanon conflict in October 2024 left a devastating impact on the country, particularly in the southern regions. The destruction of homes, infrastructure, and public services created an urgent need for accurate data to guide recovery efforts. In response, the Humanitarian OpenStreetMap Team (HOTOSM) launched an ambitious project to map the affected areas, with volunteers mapping approximately 200,000 buildings in the South Lebanon and Nabatieh governorates over the past six months. This effort has been critical in providing a foundation for reconstruction and resilience-building in the region.

The Scale of Destruction and the Need for Mapping

The conflict caused widespread damage across Lebanon, but the southern regions bore the brunt of the destruction. Researchers Corey Scher and Jamon Van Doek found that the estimated total number of likely damaged or destroyed buildings in Lebanon is about 18,500 buildings since 26 September 2023. 

The damage has major consequences to people displaced by the conflict to be able to return to their homes, and for Lebanon to recover. According to the World Bank’s Rapid Damage and Needs Assessment (RDNA) conducted in March 2025, over 30% of residential buildings in South Lebanon were damaged, with many completely destroyed. The report estimates that $1.5 billion will be required for recovery efforts, with a significant portion allocated to housing reconstruction and infrastructure rehabilitation. The mapping efforts by HOTOSM have been instrumental in identifying the most affected areas, enabling aid organizations and government agencies to prioritize resources effectively.

The RDNA report also highlights the damage to critical infrastructure, including schools, hospitals, and water systems. For example, 15% of schools in the southern regions were rendered unusable, disrupting education for thousands of children. The mapping data has been crucial in planning the reconstruction of these facilities, ensuring that communities can regain access to essential service and people can return home permanently

Collaborative Mapping Efforts

To tackle the immense task of mapping the damaged areas, HOTOSM divided the work into 11 distinct projects, each focusing on specific regions within South Lebanon and Nabatieh. This structured approach allowed volunteers to systematically cover the affected areas, ensuring that no community was left behind.

One of the key initiatives in this effort was the organization of mapathons, collaborative events where volunteers come together to map specific areas. The first mapathon was held in January 2025 in coordination with the Lebanon Reforestation Initiative. This event brought together local and international volunteers, who worked tirelessly to map thousands of buildings in a single day. The success of this event demonstrated the power of community-driven initiatives in addressing large-scale challenges. We are thrilled to see growing interest from various institutions, including the Lebanon Reforestation Initiative and Reach Initiative, who have expressed interest in collaborating with us. Additionally, educational institutions like Balamand University and American University of Beirut (Beirut Urban Lab) have invited us to their City Debates event in April 2025 and are interested in hosting a mapathon with us. Furthermore, the Ministry of Culture has invited us to organize a mapathon focused on heritage sites in Lebanon during April 2025. These opportunities highlight the potential for continued collaboration and impact through mapping and data-driven initiatives.

Building Local Capacity for Sustainable Recovery

After the initial mapathon in Lebanon, HOTOSM organized a Training of Trainers (ToT) session for experienced GIS professionals in February 2025. This initiative aimed to equip participants with the tools and methodologies needed to use the HOTOSM platform to ensure the highest of data quality. The trained professionals are now leading mapping efforts and conducting additional training sessions, ensuring that local capacity for mapping and data analysis continues to grow. This focus on capacity building is essential for creating sustainable recovery solutions that empower local communities.

The RDNA report emphasizes the importance of community engagement and local capacity building in the recovery process. By training local professionals and involving community members in OSM and other mapping efforts, HOTOSM is ensuring that the recovery process is inclusive and responsive to the needs of those most affected by the conflict.

The Road Ahead: From Rubble to Resilience

The mapping efforts in South Lebanon are more than just a technical exercise; they are a vital step toward rebuilding lives and communities. The data collected by HOTOSM volunteers is being used to guide reconstruction projects, prioritize aid distribution, and plan for long-term resilience. For example, the mapping data has already been used to identify areas where housing reconstruction is most urgently needed, helping to ensure that displaced families can return to safe and secure homes.

The RDNA report also highlights the importance of disaster risk reduction in the recovery process. By incorporating resilience-building measures into reconstruction projects, such as strengthening buildings to withstand future conflicts or natural disasters, Lebanon can reduce the risk of similar devastation in the future.

A Call to Action

The progress made so far is a testament to the dedication of HOTOSM volunteers and the power of collaborative efforts. However, much work remains to be done. The RDNA report estimates that 50,000 homes still need to be rebuilt, and thousands of families remain displaced. Continued support for mapping and reconstruction efforts is essential to ensure that all communities in South Lebanon can recover and thrive.

The below map provides a visual representation of the areas covered in the mapping projects, highlighting the scale of the efforts and the progress made so far. It serves as a reminder of the challenges that remain and the importance of continued collaboration in the recovery process. 

If you are interested in volunteering, partnering, or supporting our work in Lebanon, contact data@hotosm.org or visit HOT’s Tasking Manager projects in Lebanon. 

Humanitarian OpenStreetMap Team’s work in conflict affected areas

This work in Lebanon, sponsored by the H2H Network, is part of HOT’s Conflict & Displacement Program. In communities affected by conflict, remote mapping in OSM allows for safe and efficient data collection in hard to reach areas.  As the program grows, HOT aims to connect OSM Lebanon and similar communities affected by war to open source tools, and a global network of OSM, humanitarian, and data experts for a better informed recovery. Read more here:  hotosm.org/programs/conflict-displacement.

Said Abou Kharroub is a Local Consultant with Humanitarian OpenStreetMap (HOT) for Lebanon.

---

Dr. Samantha Lakin, Curriculum Specialist for the Organizational Learning Unit in the Office of Planning, Policy, and Resources within the Department of Public Diplomacy and Public Affairs at the U.S. State Department, sat down with Sophie Hayes from the Signal Program to discuss her background, her work at the U.S. State Department, and her goals for collaboration with the Signal Program’s current project: the Atrocity Prevention Lab (APL). Dr. Lakin is a newly added member to the APL’s Steering Committee, a group of experts convened to guide the APL’s community of practice as it serves as a dynamic platform for knowledge exchange, research, and practical interventions.

Could you share a little about yourself, your background, where you are currently, and what led you into this space?

I have focused on atrocity prevention and early warning for more than 15 years in several capacities. First, I hold a PhD in History from the Strassler Center for Holocaust and Genocide Studies at Clark University. My doctoral research explored questions of memory and justice in the aftermath of genocide in Rwanda, specifically documenting ordinary citizens’ diverse experiences with genocide remembrance processes. I have spent more than 10 years in Rwanda as Fulbright Scholar and doctoral researcher. I worked at the Genocide Archive of Rwanda and as a Policy Officer for Aegis Trust and the Kigali Genocide Memorial. I was also a Community Consultant for the preservation of the Nyamata Genocide Memorial through the U.S. Ambassador’s Fund for Cultural Preservation. We were the first team to preserve a genocide memorial site as cultural heritage. In 2024, Nyamata was designated a UNESCO World Heritage Site.

I consider myself a comparativist, with a deep specialization in the Great Lakes Region of Africa. I speak French bilingually and have a working knowledge of Kinyarwanda, Rwanda’s national language, which I was fortunate to study at Harvard and Clark.

I also have a Master of Arts in International Law and Diplomacy from the Fletcher School at Tufts University. The reason why I mention both is because when I started my PhD at Clark, I already had work experience and a skill set that allowed me to take consulting positions in post-conflict peacebuilding and human security.

My entrance to this space was, like many of my colleagues, not straightforward. After college I worked for 4 years before going back for my Master’s. I did Teach for America in Greater New Orleans, which was amazing, but very challenging. It made me realize I did not want to be an elementary school classroom teacher.

I then went on to do a Fulbright in Switzerland, where I looked at the case of Jewish children who were clandestinely rescued from France to Switzerland to escape Nazi persecution. This research explored a historical case of international rescue, which connects to my current work, interestingly enough. This work made me realize that I really like research. I liked knowing and understanding people’s stories and trying to figure out the human aspect of problems that are usually looked at with different types of data.

At Fletcher, it was harder to connect the case of rescue during the Holocaust to more contemporary issues. I owe my introduction to the field of transitional justice to Dr. Eileen Babbitt, a well-known conflict resolution professor at Fletcher School. She opened my eyes to a field that encompassed and embraced the kinds of questions I had been asking. While at Clark for my doctoral research, I applied my background from my Master’s to take direct consulting work in human security and international justice across the peace building and atrocity prevention world. I really enjoyed the duality that allowed me to develop different skills and remain relevant outside of a traditional academic path. I met a lot of wonderful people along the way that have really enhanced my work and perspective.

I learned about the APL when I was a Lecturer at UMass Boston in their conflict resolution program, which is how I met Mads [Madhawa Palihapitiya, Early Warning Consultant for the APL].

Currently, I work in the Department of Public Diplomacy and Public Affairs at the US State Department, in the Office of Planning, Policy, and Resources (R/PPR). More specifically, I work in the Organizational Learning unit of R/PPR. Our team provides support and policy guidance to approximately 5,000 people working in U.S. embassies abroad in public diplomacy roles.

I also live in Boston with my husband and our two wonderful children.

Can you speak more to your work in the federal government?

In the Organizational Learning Unit, my portfolio focuses on developing the Public Diplomacy (PD) Framework Competencies project, in addition to leading workshops on design thinking for U.S. Missions around the world. The PD Competencies project delves deep into longer-term goals of skills development, critical thinking, and assessing core competency retention across an entire PD post. For example, if an American public affairs officer rotates every three to four years, we must examine how U.S. Missions sustain high-level competencies and strengthen areas for growth among incoming officers and locally employed staff who remain in their positions.

Our unit’s work helps public diplomacy practitioners think and act more strategically, aligning programs and initiatives with US foreign policy goals. Public diplomats have a huge impact on security, and understanding national interests is essential to their work.

The goal is to really introduce a shift in mindset within the field of public diplomacy. We want to encourage the idea that public diplomats are not just event photographers, but play a hugely strategic role in carrying out US foreign policy and dealing with all the challenges in the world today. There’s a lot of different ways that looks, but it can be summed up by examining best practices as well as competencies.

You also have a long-standing connection to Harvard University, in many different areas. Can you speak a little to that?

Yes! I have been contracting and working on different initiatives at Harvard for a while. I had a fellowship in the Program on Negotiation (PON) at Harvard Law School (HLS) during the fifth year of my PhD. This was specifically a dissertation completion fellowship, but I was able to be involved in the HLS and PON community. More recently I was asked to serve as a research consultant for the Institutional Antiracism and Accountability (IARA) project on their “Global Justice, Truth-telling and Healing” research. I worked closely with Dr. Khalil Gibran Muhammad there.

So, I’ve always had these connections to Harvard.

Yes, I was going to ask about your time in the Program on Negotiation at Harvard Law. While transitional justice and negotiation studies look at post-conflict relationships, the APL is more focused on atrocity prevention and intervention-based early warning mechanisms. I’m interested in the connections you see between these two practices, and the potential applications of similar or different strategies within these spheres, having worked in both.

Post-conflict negotiation is generally cast as a relatively formal process. Best practices in international mediation focus on questions such as, “Who is the best mediator?”; “Is the process impartial?”; “How do we get people to come together and compromise?”; “How do we frame compromise?”; and “How do we deal with sensitive topics in the presence of former perpetrators?” These are all important questions necessary to create the best outcome for a negotiated settlement. However, within this framework, I believe that many of the ways transitional justice maps onto stakeholder negotiations haven’t been appropriately examined or applied. If you reframe or recast core transitional justice processes alongside negotiation practices, it can create more opportunities regarding the form and function of what justice, compromise, and post-conflict rebuilding might look like. I think it would be very beneficial for more literature to emerge that examines how these fields inform each other.

With regards to atrocity prevention, the cycle of extreme violence always stems from something. We often see a relapse or recurrence of violence when there was a history of violence that was never addressed. Prevention will always be intertwined with the past. If you don’t find ways for societies to come to terms with historical injustices, new marginalizations and social divisions will often be created based on grievances that were never addressed. People may feel like they are not in a place to create a good life, or that they haven’t been included in negotiations or guaranteed non-repetition. In these cases, there emerges a risk of more conflict. I see dealing with past crimes as a central component of prevention work.

Prevention will always be intertwined with the past.

Atrocity prevention and early warning are, to me, two of the most challenging and elusive topics for several reasons.

First, one main and popular approach to atrocity prevention often focuses on measuring different early warning factors with the goal of making an evaluation that says, “This is a situation ripe for conflict.” The question for me is, and remains, what is the right ‘cocktail’ of early warning indicators? What is the tipping point? What is the thing that essentially blows the keg? In Rwanda, for example, the catalyst for the 1994 Genocide against the Tutsi was the plane that was shot down [carrying Rwanda’s then-president, Juvénal Habyarimana]. Relying on quantitative indicators is very challenging because we can’t hold every variable constant in order to isolate and determine the impact of one variable indicator We also can’t necessarily standardize them across countries or cases. This makes any kind of atrocity prevention or early warning indicator-based system a challenge. To be sure, many quantitative researchers use these methods and do them quite well. Yet, I find that they will always be imperfect.

The second issue is that there are simply things that are just not measurable, or, at least, accurately measurable. As social or political scientists, we can try to measure social concepts like “social cohesion” or the “perceived impacts of public apologies” but we cannot always measure them in a way that translates meaningfully. There will always be endogenous variables. But people like data. People like numbers and models. The truth is that there are always going to be things that we can’t see and that we can’t measure. Trying to measure can sometimes be an inappropriate way to answer a question. Sometimes it’s necessary for scalability, but if you are doing a household survey versus talking to people on the ground about what is going on, you are more likely to have disconnect and get it wrong.

The truth is that there are always going to be things that we can’t see and that we can’t measure.

The third issue, for me, is that we don’t know what we’re preventing until it happens. We need to learn more about the null or “non-cases,” where genocide could have happened, but it didn’t. How do we work these types of cases into prevention models?

There is a lot of great development of models and measures currently in the field. In my experience, I am concerned that none of them, or at least, none of them on their own, can handle the task of comprehensive prevention. This is a hard thing to sit with. And yet, we still do this work.

Speaking specifically about the Atrocity Prevention Lab and the spatial aspect of atrocity prevention and early warning, from what I’ve seen on the ground, people are very concerned about being protected in chaotic conflict situations. This fear ends up fueling a significant amount of citizen and participatory data collection. In 2015, when I worked with Burundian activists in North America after the the third election [of Pierre Nkurunziza, who was elected to his third presidential term in 2015], people were documenting crimes with their phones and on Twitter, but without a repository or a safe mechanism to transfer and store this data for analysis. We kept telling people to document everything as evidence for the possibility that there would be a trial or case against Nkurunziza. It feels insufficient to tell people to ask people to document crimes for future, not immediate, justice. Unfortunately, it’s often the truth.

In Syria, people have been documenting crimes for years. Beginning in 2011, the Impartial and Independent Mechanism (IIIM) has collected, documented, preserved, and analyzed evidence of human rights violations in Syria under the Assad regime. The IIIM has received robust sponsorship from the international community. But the goal is the same, document now to use as evidence later.

I think the data collection aspect is an area where emerging technology and spatial analysis can really contribute. We now have an added issue of misinformation and disinformation, which include deepfakes, doctored images, and even presidential speeches fabricated by AI. It’s important to have forensically information in such a time as this, when there is pushback on real information. Forensic truth can say that such a thing happened at this time, in this location, and between these people. My work often deals with multiple truths: forensic truth and narrative truth can, at the same time, be different and valid. But when practitioners fail to differentiate between forensic truth and accurate information, we risk exacerbating disinformation.

I am excited to be a part of the APL to contribute my foundational and nuanced knowledge of the history, theory, practice, and initiatives about atrocity prevention to a new community of practitioners who have technical knowledge and expertise in geospatial technologies and spatial mapping. I know the capacity for this work and what is possible. I’m looking forward to bringing my research and practical perspectives and deep experience in the atrocity prevention world to think collaboratively and deeply about spatial aspects of atrocity prevention a little differently. There’s real room to innovate.

Further, I see a natural connection between the geospatial world and more traditional prevention efforts. Using GIS and mapping to accurately establish how, where, and when crimes took place plays an essential role in our ability to analyze trends and create established factual information to prevent bad actors from wielding information for negative reasons. After spending significant time in the research and policy realm, as well as in the field with many different communities, particularly victims, I believe there is incredible power in emerging spatial technologies. There are people already doing it, but I think the APL can do it well.

I think it’s important to have you on our team as someone who is super foundational in atrocity prevention methods and transitional justice. There is a more traditional qualitative aspect of research on atrocity prevention, and integrating these methods into the geospatial space has so much potential. It can certainly be done, but it needs to be done ethically, and that’s where your expertise comes in. How do you approach ethics in your work, and how do you feel these considerations can be integrated into your future work with APL?

Everything to me is research. Even program implementation requires elements of research. Figuring out the goals of a program, what we want to achieve, and the ways in which we can mitigate harm require [ethical] awareness. I take these practices very seriously, because, like many others, I have seen the impacts of unintended consequences for communities when people do not do their work ethically. There have certainly been times in my career where I have said, ‘This is not something I’m going to publish,’ or ‘This is an interview I’m not going to do.’ Specifically with regards to data and data collection, I question, ‘Who is this information for?’ and most especially, ‘Is this information going to help anybody?’

These questions are paramount to me when dealing with both qualitative and quantitative data. Everything comes from people. Social media, tweets, it all comes from people. People are literally putting their lives in danger to potentially get information to the outside world. It is important to evaluate the risks and benefits of this sort of data collection, and to explore how spatial technology and methods might increase safety and reduce risk.

First, thinking through how we can make this process feel like a safer and more valuable practice for people is important. The ways in which we communicate the goals of this data collection can help people to feel like their efforts are truly important, even if they don’t see an immediate result.

Secondly, this engagement can support the development of questions around what we consider as data. What do we need? What is worth it to collect? I think it was Cindy Caron [Associate Professor in Sustainability and Social Justice, Clark University] that said she never wanted to hear the term ‘data mining’ again.

I hate that phrase too!

I remember this comment so strongly, because people will use it, and it does not sit well with me. It’s frustrating to hear because that is exactly the opposite of what I want to embody. Power cuts through everything, and people’s lives are more important than anything else. When doing sensitive work, there’s always an exciting part where we, sitting in Boston or Washington, D.C., can foresee the positive impact and implications of our research or policies. But that does not mean we are ready or able to carry out prevention work without understanding the situation of our partners on the ground.

There’s always a push and pull between the policy world and the research world. People say the research world is too slow and the policy world is too fast. It’s important to find a niche where thoughtful engagement contributes to the world moving forward at any sort of pace. The APL is a unique initiative that brings together researchers, practitioners, and policy experts from diverse backgrounds, to look at the problems and potentials of spatial analysis and atrocity prevention from a multi-disciplinary and multi-directional perspective. I have high hopes for the APL and am proud to be among esteemed colleagues on the Steering Committee.

Samantha Lakin, PhD, is Curriculum Specialist for the Organizational Learning Unit in the Office of Planning, Policy, and Resources within the Department of Public Diplomacy and Public Affairs at the U.S. State Department. She is also Senior Fellow at the Center for Peace, Democracy, and Development at University of Massachusetts-Boston. Dr. Lakin’s views expressed in this article are her own as an independent academic and do not represent the views of her current employer.

Sophie Hayes is a current MS GIS student in the Graduate School of Geography at Clark University and a graduate trainee at the Harvard Humanitarian Initiative’s Signal Program.

---

Could you share a little about your background and your current work with New Lines Institute?

My background is in terrorism and counterterrorism studies. I was at a couple of think tanks before I was at New Lines Institute. I’ve run a couple of international teams at the Think Tank Civil Society Project at UPenn, a project that was led by Dr. James G. McGann, who sadly passed away November of 2021. If you look at think tank rankings, and their website says that they are ‘number one in this thematic area, or this region,’ the project defined those rankings. I led a team of 12 and then 8, with a range of expertise, from undergrads up to mid-career, and all around the world. We mapped and analyzed about 586 think tanks across MENA.

I then moved to the Institute for Study of War on their Syria and ISIS portfolios, which has become quite relevant in light of recent events.  I did some in-the-weeds research on Hayat Tahrir al-Sham, Hurras-al-Din, and the Turkistan Islamic Party- essentially the Al-Qaeda ecosystem in northwest Syria.

At New Lines Institute, I, before my current portfolio, led the Non-State Actors Program. We had a few regions of expertise: Central Asia, East Africa, West Africa, and a more broad global approach. This global approach was more geospatial in terms of looking at actors across a landscape. It was very fact-driven: we had a landscape report approach to that. We would do our roundtables, our policy engagements, but these were always after we had done a deep dive into the landscape and the actors operating there.

And I’ve taken that approach and fine-tuned it to a research methodology and a model that I use for all of the work for the portfolio that I currently lead, which is called ‘Priority Sustainable Counterterrorism.’ It’s essentially prevention, but the portfolio is named intentionally for the audience that I’m trying to engage and inform, that being more traditional security policymakers. The intent of all of the engagements and the research that we do is to bring evidence and mapping and local insights to these conversations. The goal is essentially to professionalize the conversation that sits at the intersection of security and development.

This is something that is usually a qualitative conversation in terms of research approaches. How do humans act? How do individuals engage with these groups? It’s a lot of psycho-social analysis, and a lot of things that usually fit in the qualitative realm, but there are ways to map and to bring evidence and data into those conversations. And this reveals a huge gap in the space, one that is hugely necessary to fill, and one that can bring light to any practitioner or policymaker that doesn’t see how something like supporting civil society has anything to do with conflict mitigation.

Saying, ‘Trust us and listen to the story’ is part of it, but there also has to be some tone of evidence to inspire action. So we’ve been leveraging that kind of work in a cross-sector, cross-level approach in terms of who it is that we’re engaging, and then also leveraging multiple facets in the work that we do. We leverage web sources, OS-INT (open source intelligence), academic literature reviews, but also dig into existing NGOs and conflict think tanks. This final process looks into the reports that such organizations are producing, as well as the local consultations and databases that already exist and can be leveraged into the work that we do. Basically, the portfolio is set up to collect as much information as possible, and then bridge existing gaps.

Geospatial can also be an application; a fact-finding mission. Rather than a tool, it can be a form of work.

When you approach the integration of psycho-social analysis and the qualitative realm with geospatial analysis, what sorts of tools do you use?

What I am piloting, and what I hope to see others pilot as well, is the idea of geospatial analysis as a methodology past a really ‘in the weeds’ tool. I think more terrorism, conflict prevention, and peacebuilders can engage in this process. When you talk about geospatial tools and GIS, it can be scary to someone that’s not trained in it. There’s only so many people that know how to do the coding and the analysis to produce these products. Geospatial can also be an application; a fact-finding mission. Rather than a tool, it can be a form of work. Whether you’re looking at SIPRI (Stockholm International Peace Research Institute), military equipment sales, or ACLED conflict event data, you’re looking at geospatial data, and these data can be used to make portfolios more robust without applying technical skills.

For my portfolio, I produced two reports about the intersection of security and development policy in the Sahel, and integrated location-based data with qualitative research. I did market research and mapped over 300 civil society organizations across the Sahel, following this up with 10 in-person visits with NGOs at their place of work, and another 60 online consultations with local experts and civil society leaders. These visits were designed based on their location and their size. The research model was a geospatial application of research parameters; essentially, taking a map and focusing on specific areas of conflict that intersect with civil society. The second report was designed for the security sector and produced a conflict analysis on the operations of the Al-Qaeda affiliate in the area. It found that they had a civilian engagement strategy that predated their territorial expansion.

So when I move forward with briefs for policymakers, and I recommend a comprehensive prevention strategy, I’m recommending that in order for U.S. engagement to be more effective, our development and security efforts have to be at the very least coordinated. I can show them two maps, with two applications of geospatial analysis, and overlap them. These two very different buckets of data are able to support my recommendation due to their overlap.

People ask, ‘Where’s the proof?’ They want to know if we’re actually preventing anything at all with these ten-year development strategies. And that’s a fair question. We have to get better at mapping the non-kinetic. Whether that’s the operations of a bad actor, or if that’s mapping the solution landscape, applying a geospatial lens as a form of analysis is a way to professionalize a lot of these very conceptual conversations. And that is a really big opportunity.

Is your audience for these conversations mostly policymakers?

Yes, that is who New Lines Institute usually engages with. We brief policymakers from different parliaments around the world as well as domestically, from the State Department, to the Hill, to the United Nations, and everywhere in between. The portfolio that I run engages those same policymakers in addition to practitioners and local actors, and I weigh them the same. I don’t hold policymakers to a higher weight than I would practitioners because in my opinion, policy implementation is almost more important than policy itself. I also believe that positive change is not only iterative but also a shared responsibility. While we may have some decision makers that lead the charge in terms of policy priority, or, for example, national government funding, I don’t think that them having that responsibility lowers the potential of every single person that is engaged in this space and his/her potential impact.

I do these data-heavy mapping reports for policymakers’ eyes, but in that process, I am essentially fact-checking myself against civil society. I host short-term working groups for civil society organizations, which is a separate project. I don’t mention terrorism or security in those conversations. Rather, they’re focused on localization, policy, and how governments and donors can better engage with youth-led or civil society organizations. While policymakers at the highest levels might be the main audience, I think that how we do our work is just as important. I feel passionately that putting strong attention into one meeting with a civil society organization that has a team of just 5 people in Sahel, or Kenya, or wherever it is that I’m working, can have just as much impact as a briefing at, say, the Office of Counterterrorism.

This is also where geospatial as a framework comes in. Because whether we refer to geospatial as a tool, or technology, or a form of analysis, we’re talking about data. At the smallest level, on a map that has many countries, or many towns, each individual node is implementation. And implementation is all about people. Counterterrorism is all about events that impact people. And so, at the smallest level, the people who we’re engaging with should also be informing and guiding us, or be part of what we’re trying to achieve.

As you know, the Signal Program itself was started as a more traditional conflict surveillance mechanism, and in the past several years has switched almost completely to working with end-users and practitioners. It’s so important that you can look at the differences between these two groups and bring expertise to interacting with both of them. How do you feel your conversations with policymakers versus practitioners are different, and what kinds of things do you weigh more heavily when conducting these conversations?

Every single actor we engage with has their own language and their own priorities. This is really important to consider. Whether I’m developing a partnership with an international NGO or I’m having a meeting with a civil society organization that might not produce a formal partnership, there is information sharing present in both of these relationships, and I have a responsibility to give something back to them. This can simply be a connection, or insight of my own.

With policymakers, although it varies between each departmental office, I have to understand who the person I’m speaking with is beholden to. I have to understand their parameters, priorities, and most importantly, what they can and can’t do. I then have to look at what I am briefing them on, and find the connection point, which is something actionable. That, to me, is what impact is. Impact is studying something and saying something in a way that is able to be heard. A lot of times, when we do work and it lands on ears that don’t understand, or don’t see its relevance, this work does not become impactful in any way.

The main goal is to change the world. To address, mitigate, and prevent conflict, we need data as a bridge builder.

How do you work with more data-heavy information within these contexts?

Especially on the tech side, things are often hard to translate. It might feel like it isn’t relatable, or it can’t be translated to a more general audience. But there is always a way to have a key takeaway, or to have an insight that translates, even if sometimes you have to take two steps back instead of just one.

Geospatial data is so much more relevant than we give it credit for. There are so many different spaces in this world of prevention, whether we’re between academics, practitioners, or peacebuilders, or between the humanitarian, human rights, and conflict resolution sectors. They have an overarching, shared objective, and some of them may share ideas on how to fix problems, but data is the bridge builder between all of these spaces. I see so much opportunity for the Atrocity Prevention Lab (APL), and for geospatial analysis and data-based approaches in general, to resolve so many of the disconnects that we see today.

The main goal is to change the world. To address, mitigate, and prevent conflict, we need data as a bridge builder.

To conclude, can we hear more from you about your intentions for the Steering Committee, and what you’re excited to bring?

I just see so much opportunity. There has been so much progression in all of these different sectors, from terrorism and counterterrorism, to peacebuilding, to negotiation and mediation. Every single sector and sub-sector has waves of growth and change, and they are all trying to be relevant as discussion and work spaces.

APL is holding on to so many data-informed reasons that things should be done the right way, and it is my mission to show these reasons to the people that are still deciding.

I think there are really just a handful of relevant spaces right now that are at the point of realizing they’re intersecting with other silos. And who knows what brought that on. Maybe it was brought on by us being better connected, like the fact that I can hop on a call right now with someone from Kenya, and then fifteen minutes later I can have a call with someone in Nigeria. How amazing is that? But I think that all these spaces need a general umbrella of addressing and preventing violent events and conflict. Whether that’s positive peace, as the Institute of Economics and Peace has a wonderful framework for, or addressing more kinetic problems, such as the operations of an Al-Qaeda group, I think a common language across so many sectors is data. And right now, we’re all speaking it in different ways.

That’s why I think APL has a really unique opportunity right now to connect so many efforts of so many different entities across the world, on such an important topic. And most importantly, it’s a partner effort. APL isn’t the only one doing amazing things, but being a part of the steering committee, I’m really excited to see how we can be leaders in this space. That’s what everyone is looking for right now.

All the great work that APL and [the] Signal Program has already done is quite clear. So I would like my contribution to be a focus on bringing this work to policymakers. While it’s a really well-done academic and practitioner-focused project, there’s so much that can be gleaned to help guide policy and policy implementation. APL is holding on to so many data-informed reasons that things should be done the right way, and it is my mission to show these reasons to the people that are still deciding. We might look at our work as so obvious, but the fact of the matter is that there are a lot of decision makers around the world that don’t see it. We have this bucket of reasons to prioritize sustainable approaches to mitigating and preventing conflict, and this is very much in line with my objectives at New Lines.

Tammy Lynn Palacios is Senior Analyst & Portfolio Manager of the Priority Sustainable Counterterrorism portfolio at New Lines Institute for Strategy and Policy. She is also a 2023-2024 Modern War Institute Research Fellow.

Sophie Hayes is a current MS GIS student in the Graduate School of Geography at Clark University and a graduate trainee at the Harvard Humanitarian Initiative’s Signal Program.