What happens to my data when I use free AI writing tools?

When you use free AI writing tools, your inputs — the prompts you type and the text you paste — are typically stored, reviewed by human employees, and used to train future versions of the AI, which means anything confidential or personal you share isn't really private. This isn't a secret. It's right there in the terms of service most people click through without reading. OpenAI, for example, has stated that conversations from free ChatGPT users may be used for training unless you specifically opt out through their privacy settings. Google's Gemini has similar policies for free-tier users. I've talked to small business owners who were pasting client contracts into free AI tools to get summaries, completely unaware those documents were being logged. That's a confidentiality nightmare. The data flow works roughly like this: your text goes to the company's servers, gets processed, and the interaction may be sampled for quality review or model improvement. Even if the data is 'anonymized,' research from Princeton and other institutions has shown that training data can sometimes be extracted from language models through clever prompting. The practical takeaway: never put anything into a free AI tool that you wouldn't put on a public whiteboard. No client names, no financial figures, no unpublished manuscripts. Paid business plans from major providers usually offer data processing agreements and promise not to train on your inputs — but always verify this in the specific plan's terms. For a deeper dive, see our guide on AI content copyright and legal issues. **Related**: Are paid AI tools more private than free ones? | Can AI companies read my private conversations?