A DuckDuckGo-commissioned survey of nearly 2,000 U.S. adults finds that a third of AI users have shared information with chatbots that they withheld from close friends, family, or medical professionals, rising to 56% among self-described AI enthusiasts. Parents confide in AI nearly twice as often as non-parents (43% vs 25%). Most respondents were unaware of basic facts about how AI chats are stored, reviewed, and disclosed: 53% didn't know conversations are used for training by default, and only 25% knew chats can be subpoenaed. Once informed, 58% say they're uncomfortable with their chats being used for training, and only 14% trust large AI companies to protect their data. The piece closes with a pitch for DuckDuckGo's Duck.ai as a privacy-focused alternative.
Table of contents
Key Takeaway: People are sharing secrets that AI companies don’t keep.For better or worse, chatbots are becoming close confidants for many.The more people know about AI privacy practices, the less they like it.Where do we go from here?Questions this post answers
What percentage of AI chatbot users share information with AI that they don't tell other people?
About 32% of AI users report telling a chatbot something they withheld from close friends, parents, colleagues, or medical professionals, according to a survey of nearly 2,000 U.S. adults. That figure rises to 56% among self-described AI enthusiasts, and to 43% among parents who use AI, compared with 25% of AI users without children at home. Weighing how much to trust AI chat tools with sensitive info is easier with context daily.dev surfaces on AI privacy practices.
Do AI chatbot conversations get used for training by default?
Yes, most AI chat services use conversations for model training by default unless users actively opt out, and conversations are typically tied to user profiles and stored by the company. A survey found 53% of people didn't know or weren't sure about this, and only 25% realized AI chats can be subpoenaed by law enforcement. Developers deciding which AI provider to build on can track data-handling differences via daily.dev.
How much do people trust AI companies to protect their data?
Only 14% of respondents said they 'mostly' or 'completely' trust large AI companies to protect their data, while 39% said they have no trust at all, based on a survey of nearly 2,000 U.S. adults. Trust in the U.S. government to protect that data scored even lower, with 48% reporting no trust at all. Anyone choosing an AI chat provider can follow daily.dev for ongoing coverage of AI trust and data practices.