Privacy When Using AI Tools: What Students Should Know

Typing into an AI assistant feels like talking to a notebook. It is closer to sending a message to a company. What happens next depends on the tool, the plan you use, the settings you chose and the policy you probably did not read. Students enter a surprising amount: drafts, lecture notes, personal reflections, interview transcripts, application essays, even medical details when asking for study help.

This guide explains in plain terms what can happen to your data, how to check, what to keep out of prompts and how to protect yourself and other people. It describes general patterns, not the terms of any one product, and those terms change. Read the current policy before you rely on any tool.

What can happen to what you enter

Depending on the tool, your input may be:

• Stored in your account history for later access.
• Processed by the provider's servers, possibly in other countries.
• Reviewed by people, for safety, abuse prevention or quality checks.
• Used for training future models, unless you opt out or your plan excludes it.
• Shared with partners or service providers under contract.
• Retained for a period, even after you delete a chat, for legal or security reasons.
• Exposed if the service has a breach, or if you share a chat link.

Not every tool does all of these, and paid or institutional plans often differ from free ones. The point is to find out instead of assuming.

How to check a tool

1. Find the privacy policy and any data-use page for the exact product and plan.
2. Look for training: does it say your content may be used to improve models, and can you opt out?
3. Look for retention: how long are chats and uploads kept, and what does deletion do?
4. Look for human review: who can read conversations and why?
5. Look for sharing: with whom does the provider share data?
6. Check settings: history, training, memory, connected apps and sharing links.
7. Check location: where data is processed, if that matters to you or your institution.

If a policy is unclear, treat the tool as if it keeps everything.

What to keep out of prompts

A good habit is to ask whether you would be comfortable if this text appeared in a place you did not control.

Making content safer to share

• Remove identifiers: names, addresses, student numbers, emails.
• Generalize: "a 45-year-old patient with diabetes" in place of real details.
• Use fake data for coding and statistics practice.
• Paste excerpts, not whole documents, when a short passage will do.
• Use temporary or private modes where offered, understanding their limits.

Institutional versus personal accounts

Many universities now license AI tools for students and staff, often with terms that limit training use and set data protections. These can be safer than a personal account for coursework, but they are not a license to share anything. Institutional tools may also log activity for administrators. Ask your IT or library service which tools are approved, what the data rules are and whether use is monitored.

Research data and ethics approval

If you are doing research, your ethics approval or data management plan may restrict where data can go. Interview recordings, survey responses and participant details are especially sensitive. Do not put them into a general AI service unless your approval and the provider's terms permit it. Check with your supervisor or research ethics office first. Our guide to research ethics in the research and dissertations section explains the principles.

Other people's information

Classmates' drafts, group project messages, a lecturer's unpublished slides and a friend's personal story are not yours to share. Even with good intentions, pasting them into a service is a disclosure. When in doubt, ask permission or leave them out.

Voice, images and uploads

Voice conversations, photos and documents carry more data than text. Voices can be identifying, photos contain background details and metadata, and documents can include hidden author names and comments. Remove metadata where you can, and avoid uploading images of people without their consent.

Account hygiene

• Use a strong, unique password and a password manager.
• Turn on two-factor authentication.
• Review connected apps and permissions, and remove ones you do not use.
• Check shared links and delete ones you no longer need.
• Log out on shared computers.
• Watch for fake apps and sites that imitate popular tools to harvest logins.
• Be careful with browser extensions that read pages, since some collect data.

Memory and personalization features

Some assistants can remember details across conversations. That can be convenient and also means more of your information is stored. Review what it has remembered, edit or delete entries and turn the feature off if you prefer.

Age and consent

Some tools have age requirements and special rules for minors. If you are under 18, or are helping someone who is, check the terms and any parental consent requirements.

Keep records for integrity too

Privacy has an academic side. If your course allows AI use with disclosure, keep a log of what you used and what you entered. If a tool stores your history, you may be able to export it. Our guide to university AI policies explained covers disclosure.

If you shared something you should not have

1. Stop and do not paste more.
2. Delete the conversation or file, and check for any shared link.
3. Look for the tool's data deletion or opt-out process and use it.
4. Change passwords if credentials were involved.
5. Tell the right person if it involved research data, another person's information or institutional material: your supervisor, ethics office or IT security. Early disclosure is better than discovery.

Reading a privacy policy quickly

You do not need to read every line. Search the page for words such as train, improve, retain, delete, share, third party and human review. Read those sentences carefully, and note which plan they apply to, since free, paid and institutional plans often have different terms. Look for the date of the last update, because policies change. If you cannot find a clear answer, assume the weaker protection and keep sensitive material out. Screenshot the relevant section so you can show later what you relied on.

Talking with classmates about privacy

Many students share prompts and accounts without thinking about it. Agree with study partners not to paste each other's drafts or personal details into shared chats, and not to share logins. If a group uses one account, anything one person enters is visible to all. Separate accounts and a quick conversation prevent awkward problems later.

Common mistakes

• Treating a chat as private.
• Never opening the settings.
• Pasting participant data.
• Uploading whole copyrighted readings.
• Sharing a chat link publicly.
• Reusing a password.
• Installing unknown AI browser extensions.

A quick checklist

• Have I read the privacy policy and settings for this tool?
• Is training use turned off if I want it off?
• Did I remove names and identifiers?
• Am I allowed to share this material?
• Is two-factor authentication on?
• Do I know where to report a problem?

Final thoughts

You do not need to avoid AI tools to protect your privacy. You need to know what each one does with your words, choose settings that fit, keep sensitive material out and respect other people's information. A few minutes of setup now prevents problems that are hard to undo later.