Understand What AI Remembers
Think of a public AI chatbot as a conversation that is being recorded and potentially reviewed. Anything you input—from simple questions to chunks of an essay—can be stored and used by the company to train its models. This includes not just your prompts
but also metadata like your IP address, location, and device information. Many popular AI services state in their privacy policies that this data can be retained and analyzed by employees or third-party contractors. This means sensitive information about your academic performance, personal life, or even unpublished research could become part of the AI's training data, potentially exposed in future responses or a data breach.
Treat AI Like a Public Square
The simplest rule for AI privacy is this: don't enter any information you wouldn't feel comfortable posting on a public forum. This includes obvious personal identifiers like your name, address, or student ID number, but also more subtle details. Avoid pasting in financial data, health information, or confidential details about your family or work. A series of seemingly innocent queries can be combined to paint a detailed picture of your life, so it's wise to treat every interaction as if it could be linked back to you, even if your name isn't directly attached. Consider any information you provide to a public AI as potentially public.
Anonymise Your Academic Work
Using AI to get feedback on an essay or solve a problem set is a popular study technique. However, uploading entire drafts of unpublished work carries risks. To protect your intellectual property and personal data, get into the habit of anonymising your work before you paste it in. Remove your name, your professor’s name, the course title, and any other specific identifiers. Instead of pasting a full essay about a specific historical event in your hometown, for instance, you could generalise the details. The goal is to provide enough context for the AI to be helpful without handing over a document full of identifiable data.
Use Built-In Privacy Features
Most major AI platforms now offer privacy controls, but they are often not enabled by default. Go into the settings of the AI tools you use and look for “Data Controls” or a similar privacy section. The most important setting to find is the option to prevent your conversations from being used to train the company's models. Many services, including ChatGPT, Gemini, and Claude, allow you to opt out. For particularly sensitive queries, use features like “Temporary Chat” or “Incognito mode,” which do not save your conversation history and are not used for model training.
Check Your University's Policies
Before you integrate any AI tool into your workflow, check your school's official policies. Many universities now have specific guidelines on the acceptable use of AI, and some have approved certain tools that come with enhanced, enterprise-grade privacy protections. Using an unapproved tool, especially for graded assignments, might violate academic integrity rules. Furthermore, federal laws like FERPA in the United States protect the privacy of student educational records, and you cannot be required by an instructor to use a third-party tool that the university has not vetted.
Separate Your Accounts
If possible, avoid linking your primary personal or student email account to public AI tools. Consider creating a separate email address used exclusively for AI services. This simple step helps limit the amount of personal data that can be associated with your account and reduces the information an AI provider can access if you sign in via a linked Google or social media profile. When you must create an account, protect it with a strong, unique password and enable multi-factor authentication to keep your chat history and personal data secure from unauthorised access.
