Home / AI literacy guides / AI basics
AI literacy guide · AI basics
Bias, privacy, and what not to share
Two cautions every instructor should carry into any use of these tools.
Because these tools learn from large bodies of human text, they absorb and can reproduce social biases, stereotypes, and gaps in that text. Output can be skewed in ways that are easy to miss if you are not watching for them, so review anything you will put in front of students.
Privacy is the other caution. Anything you type into a public AI tool may be stored and used by the company that runs it. Never enter student work, grades, identifying information, or anything covered by FERPA into a public tool. Treat the input box as public, not private, and strip identifying details before you paste.
You have some control over this. In most AI chat tools you can open the personal or account settings and turn off the use of your inputs to train future models. Take a minute to do this, and let students know they can do the same in their own accounts.
There is often a better option than a personal account. Most colleges and universities hold an enterprise contract with at least one AI provider that offers stronger data protections than a free or personal account, so prefer the institution-provided tool when you have one. The provider usually follows the systems your campus already uses: a campus on Google Workspace likely provides Gemini, and a campus on Microsoft likely provides Copilot. To find the specifics for where you teach, see your institution AI policy directory.