AI, being a technology versus a single tool, can be categorized in different ways. One way to think of AI is in what it’s used for.
Generators (generative AI)
These AI tools generate, or create, content. The content is generated using examples from a pool of media. This might mean that it generates texts, images, video, or audio.
Note: Many generative AI features are being integrated into tools that students use every day, such as Microsoft Word’s Copilot embed and Google Docs’ AI suggestions. It’s difficult to avoid generative AI when it’s embedded.
Analyzers
These AI tools analyze content, using data sources they have access to. Data sources may be databases of journal articles, websites, or otherwise. The analysis could be anything from offering editing suggestions (Grammarly) to suggesting related articles (embedded AI finding aids in databases). Many “research” AI tools fall into this category.
Note: you can also analyze data with generative tools like Copilot! For example, you can upload a spreadsheet into Copilot and ask for an analysis.
It is difficult to define many AI tools in just one way. Some research AI tools are definitively for analyzing (and suggesting), but even generative tools can have search functions. For that reason, it may be better to think of AI in terms of task. This is especially true given that new AI tools are released every day, and it can be impossible create policies for each of them!
Not sure what “tasks” generative AI can do? Check out Harvard’s guide to generative AI for feature lists to consider how it might apply to your subject area as well as their "Teach with Generative AI" page.