What is the Gemini application?
The Harvard Chan AI Glossary
In our classes, we often pause to explain the ideas behind AI and the language used to describe it. These posts grew out of those conversations. They’re for anyone who would like to understand AI better, with no technical background assumed.
The Gemini application is Google’s user-facing AI assistant. It provides a place for people to ask questions, work with material and use available connected features. Gemini is also the name of a model family, so it is important to distinguish the assistant product from the trained model doing part of the computation.
A model interprets information and produces responses. An application adds the interface, account features, stored material and tools around it. A tool is a software function that can retrieve information or perform an action. The exact combination available to a person depends on the product version, account and configuration.
As an illustration, a person might ask the assistant to explain a photograph of a handwritten plan. The application must supply the image to a suitable model and present the answer. If the person wants information from a connected service, that requires a separate working connection and appropriate access. Image interpretation and access to another account are different capabilities.
The Gemini application also differs from Google AI Studio, an environment for experimenting with models and developing applications, and from an API, an interface used by software to send requests directly. These can involve related models without providing identical working experiences.
Connected features should not be inferred from the brand alone. A generated sentence saying that a file was checked is not evidence that the application obtained it. For important work, inspect the source or the result of the actual operation. The distinction is between a model’s response and a verified event in the surrounding software.
When comparing assistants, try the same task with the same source material. Differences may come from the selected model, the files actually supplied or the connected tools, as well as the interface.
Choosing an assistant application is therefore partly a choice about workflow: how material is supplied, how results are reviewed and which actions can occur. The model contributes capability, while the application shapes how that capability becomes usable. The product name alone cannot establish the quality or authority of a particular answer.
Source: Google — Gemini Application Overview.
If you would like to find out more about AI and improve your leadership skills, consider the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.
AI is a fast-moving field. This explanation reflects our class discussions and related material at the time of writing. Terminology, capabilities and tools may change.
For suggested additions or corrections, please email me at tpanch@hsph.harvard.edu.
Dr. Trishan Panch
Program Director, AI in Health Care
Harvard T.H. Chan School of Public Health — Advanced Learning Academy