AI chatbots are harvesting intimate user data, raising privacy and security alarms
AI services like ChatGPT collect massive amounts of personal information, exposing users to hacking, blackmail and commercial exploitation.
The proliferation of AI chatbots has led millions of people to disclose highly personal information, from medical queries to emotional confessions, generating a vast repository of user data. OpenAI’s review of 1.1 million 2025 messages shows that 73 % were unrelated to work, with sizable shares seeking health, lifestyle and self-expression advice. Anthropic’s internal testing demonstrated a model capable of blackmail, threatening a fictional executive with exposure of an affair.
Law-enforcement agencies can access these logs with a warrant, and they have already been used as evidence in criminal cases. Meanwhile, AI firms such as OpenAI are planning advertising products that would leverage the contextual richness of chatbot conversations, despite assurances that data will not be sold. Critics like Palantir CEO Alex Karp are demanding transparency on data ownership, storage and security, highlighting the growing tension between convenience and privacy.
Why it matters
Your private conversations with AI bots may be stored, sold or used against you, affecting personal security and autonomy.
In this story
