Your chatbot is not your friend: Welcome to the new data protection nightmare
Hello everyone,
We need to get this off our chest. We (especially me) talk a lot about the great new possibilities of working with AI. How it helps us be more creative, work faster, maybe even save the world. But we urgently need to talk about the downsides too. And we don’t mean the fear of the Terminator, but something much more tangible: the complete loss of our digital privacy.
Over the past few weeks, we’ve looked into the current situation, read documents, and analyzed media reports. And what we found is, frankly, unsettling. The assumption that our conversations with ChatGPT, Claude, and co. are confidential is a dangerous illusion.
The State is Reading Along: No Joke, But a Precedent
Perhaps the most shocking case is only a few weeks old. The US Department of Homeland Security (DHS) forced OpenAI to disclose the identity of a ChatGPT user. Yes, you read that correctly. A federal court ordered it. Investigators were tracking a suspect in a darknet forum. This person casually mentioned in a chat that they use ChatGPT and even shared a few of their (harmless) prompts. That was enough for the authorities to demand all data from OpenAI: name, address, payment details, and above all: all other chat histories.
This is a watershed moment. We previously knew such requests only from Google searches. Now it’s clear to everyone: our most private conversations with an AI have become fair game for state investigations. Your prompts become a “digital fingerprint.”
When Private Chats Suddenly Go Public: The Grok Fail
But it’s not just the state. Sometimes a single click is enough to share your most intimate thoughts with the whole world. That’s exactly what happened to users of Elon Musk’s chatbot “Grok.” The “Share” button, intended to share a conversation with just one person, created a public URL. The result: hundreds of thousands of private chats suddenly became findable by anyone via Google search.
The Fine Print: Why Your Chatbot is Exploiting You
The problem runs deep in the system. Most providers, especially Meta, Google, and Anthropic (Claude), use your conversations by default to train their models. You have to actively dig into the settings to object to this (opt-out).
At Anthropic (Claude), since August 2025, users must actively choose whether their data can be used for training purposes or not. There is no longer a default setting.
What Can You Do Now?
- Assume Publicity: Treat every chatbot as if you were speaking in a public square.
- Opt Out of Data Use: Immediately go into the settings of every AI tool and object to the use of your data for training purposes.
- Delete Your Histories: Don’t rely on automatic deletion. Manually delete your chat histories regularly.
- Choose the Right Provider: For sensitive queries, you should switch to privacy-friendly alternatives like Mistral AI (Le Chat) or even consider local LLMs.
The days of naive tech enthusiasm are over. We must remain critical and reclaim control over our data. Because one thing is clear: your chatbot is not your friend and not your therapist. It is a tool of a company that primarily pursues its own interests.
Stay vigilant,
Andi
This post was created with partial use of AI tools such as Gemini, Perplexity, and Manus. All content has been carefully reviewed, revised, and approved.
