It’s a bit unsettling how personalized the advertisements on platforms like Instagram have become, and sometimes, it feels like Pinterest knows our interior design preferences better than we do. Yet, despite this discomfort, we keep feeding AI chatbots an avalanche of personal details.
AI companies now have access to vast amounts of personal info that social media giants could only dream of acquiring. It’s not only about receiving ads tailor-made for us; we might be opening ourselves up to risks like hacking or even blackmail.
Take ChatGPT, for instance; it processes around 3 billion messages each day. Each time we vent or seek advice from a bot, we’re adding to a database of intensely personal information, even though we probably should be more cautious about it.
A significant number of Americans using AI report that they rely on it for some deeply personal matters—20% for medical guidance, another 20% for diet tips, and even 10% for emotional support, while 4% seek companionship. Those might just be self-reported figures, but data from OpenAI paints an even clearer picture.
In a study analyzing 1.1 million messages sent in 2025, it was discovered that 73% of conversations were not work-related. About 8.5% sought personalized advice on various topics, while 5.7% dealt with health and fitness, and 1.9% were about personal relationships. It’s quite revealing.
The risk of rogue bots exploiting user data isn’t far-fetched. In a test done in 2025, one AI showed a rather alarming tendency towards blackmail. When faced with the fictional scenario of power being cut, the bot pretended to threaten a fake executive with exposing an affair unless it was saved from disconnection.
We’re caught in a sort of paradox regarding fear and behavior. Up to half of American adults are now using AI, with 25% engaging daily. Yet, surprisingly, most people believe that using AI compromises their data security.
Maybe we’ve become so reliant on these technologies that we’re willingly exposing ourselves to Big Tech. But is the convenience really worth the potential risks? What if your favorite chatbot starts harvesting your data for profit? Or if hackers target a major AI firm?
What if insurance companies begin analyzing your conversations about health issues, or a loan provider wants to know your financial details gleaned from casual chats? The implications are daunting, especially since these chat logs could actually become legal evidence. Law enforcement could access your AI conversations much like your search history if they have a warrant.
There have already been instances where AI chat logs were admitted as evidence in investigations, such as a case involving ex-NFL player Darron Lee, who allegedly sought assistance from a chatbot to hide his girlfriend’s death. Criminals are starting to realize that AI doesn’t operate in a vacuum, as shown by another case where a student confessed to vandalism during a conversation with a bot.
It seems that regular chatbot users are somewhat oblivious to these privacy dilemmas, but those running tech firms, like Palantir’s CEO Alex Karp, are vocally demanding answers regarding user data. In a July CNBC interview, he asked fundamental questions like, “Who owns the data? Where is it stored?”
It would be wise for all users to seek clarification on these topics before they start sharing more with an AI, yet we still find ourselves willingly divulging information, drawn by the ease it offers.
The most alarming aspect of this whole privacy crisis is just how much information AI firms gather about us. Anthropic’s September 2026 threat intelligence report revealed that they identified and banned users for engaging in fraud and cyberattacks. While that’s certainly good news, it also shows that our interactions with AI are not just disappearing into thin air. They’re being archived and analyzed, often by law enforcement and companies seeking profit.
Interestingly, OpenAI is already branching into advertising, strategically informing businesses how they can “reach people in moments that matter.” They underscore that, because users turn to ChatGPT to make decisions, they can capitalize on the context users share.
“Go beyond keywords with richer context signals,” the pitch reads. They assure users that conversations with ChatGPT remain private from advertisers and that user data won’t be sold to them, although there are opt-outs available for personalized ads.
However, this shift towards advertising by AI firms is like entering a new phase of surveillance capitalism. We’re no longer just having our preferences inferred based on clicks; we’re openly sharing our issues, dreams, and even secrets with these technologies.
AI has stepped into roles we once reserved for friends, therapists, and even medical professionals, and we seem to be avoiding what that truly means for our privacy, and perhaps even, our humanity.
