What is Grok?
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.
Grok is a name used for both an AI model family and an assistant product. It was developed by xAI. The developer API and user-facing assistant are different ways of accessing its capabilities. An API, or application programming interface, is an agreed way for software to send requests and receive responses.
The model is the trained component producing responses. The assistant is the surrounding product through which a person works. It may combine a model with search, file handling or other tools, depending on the configuration. A tool is a software function that obtains information or performs an action beyond generating text.
This distinction matters when someone says a system has current information. A model’s learned patterns and a live search result are different sources. A product may obtain up-to-date material through a tool, but the model-family name alone does not prove that a search occurred for a particular answer. The evidence and the actual action need to be inspected.
An illustrative example is asking about a newly published document. A supported application could retrieve it and supply the contents to a model. The resulting answer should still be checked against the document. Access to information and accurate interpretation are separate achievements.
Developer documentation describes particular models and available capabilities. The names, supported inputs and access arrangements can change, so this entry explains the role of Grok rather than providing a permanent feature table. An account or integration may expose only some of the possibilities described for the wider platform.
For a task involving current information, check whether a live search actually occurred. A model family’s reputation for access to current material does not show what happened in a particular response.
For a useful comparison, describe the task and complete setup. Which model ran, which sources were supplied and which tools were available? A polished assistant response may depend on all those components. Grok identifies a model and product family; the configured system and the supporting evidence determine what a particular result means.
Source: Grok — Developer Documentation.
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