Understand the Default Setting
Most popular AI chatbots from companies like OpenAI, Google, and Anthropic are free to use because your conversations can be valuable. By default, many of these platforms use the prompts you enter to train their future models. This means that any text
you paste—whether it's a draft of a personal essay, questions about a sensitive topic, or notes from a research project—could be stored and potentially reviewed by human developers. These conversations are collected to improve the AI, but it also creates a digital footprint you may not be comfortable with. The first step to protecting your privacy is understanding this basic transaction: in most cases, your data is the product.
Never Share Personally Identifiable Information
This may seem obvious, but it's the most critical rule. Treat every chat with a public AI as if it's a public forum. Never enter sensitive personal details into a prompt. This includes your full name, home address, phone number, national identification numbers, school account credentials, or any financial or health information. Even seemingly harmless details, when combined, can build a profile. A student might inadvertently paste notes containing names of classmates or details about a school project that shouldn't be public. A good habit is to read over any text you plan to paste and remove any information you wouldn't want a stranger to see.
Anonymise Your Academic Work
You can get help with an essay or dataset without giving away the entire context. Before pasting your work, replace specific names, locations, and other identifiers with generic placeholders. For example, instead of pasting an essay that mentions “my internship at XYZ Corporation in Mumbai,” you can change it to “my internship at a large tech company in a major city.” The AI will still understand the context needed to help with grammar, structure, or tone without you having to expose specific, private details about your life or work. This practice of redacting your own information is a powerful way to get the benefits of AI assistance while minimising your risk.
Use Your Privacy Settings
Major AI platforms have realised that users want more control. Most leading services now offer settings to prevent your conversations from being used for AI training. For ChatGPT, you can find this in 'Data Controls' and turn off the option to 'Improve the model for everyone'. For Google's Gemini, you need to go into your Google Account settings and pause 'Gemini Apps Activity'. While these settings are a significant step, be aware that companies may still retain your data for a short period (often up to 30 days) for safety and abuse monitoring. Using these opt-outs is the single most effective technical step you can take to protect your privacy on these platforms.
Consider Privacy-First Alternatives
As concerns about data grow, a new market of privacy-focused AI tools is emerging. Some platforms are designed from the ground up to not train on user data. Services like Proton are launching private AI assistants, and search engines like DuckDuckGo now incorporate anonymous AI chat features. For particularly sensitive tasks, you could even explore running smaller, open-source AI models locally on your own computer, though this requires more technical skill. If you are using AI for university work, check if your institution provides a license for an enterprise version of an AI tool, as these often come with stronger, contractually-obligated privacy protections.
Create Separate and Anonymous Accounts
A simple but effective strategy is to compartmentalise your AI usage. Avoid signing up for AI services with your main email account, especially one that contains your full name or is linked to your other online identities. Instead, use a separate, anonymous email address to create your AI accounts. This adds a layer of separation between your AI activity and your personal life. Furthermore, some platforms like ChatGPT allow you to use them without an account for basic queries, offering a quick way to ask a question without any of your activity being saved to a profile. This prevents the AI company from building a detailed history of your usage patterns tied to your primary identity.
