One in Three Tell AI Chatbots What They Hide From Everyone

Photo Credit: MMD Creative

A new survey from DuckDuckGo puts numbers to a habit many users may not fully register: nearly one in three people have told an AI chatbot something they would not share with friends, family, colleagues, or medical professionals. Among self-described AI enthusiasts, that figure climbs to 56%, according to the private search company.

The survey drew responses from nearly 2,000 U.S. adults and focused on the gap between how people use AI chatbots and how much they understand about where those conversations go. DuckDuckGo argues that the format itself invites disclosure. Where a typical web search runs a few words, chatbot exchanges pull longer, more personal input from users. In the company’s framing, search queries reveal interests, while AI conversations “have the potential to reveal thought processes, communication styles, and more.”

Much of that sharing happens without a clear sense of what follows. DuckDuckGo found that 53% of respondents did not know or were unsure that most chatbots use conversations for training by default, and just 25% knew that AI chats can be subpoenaed by law enforcement. Awareness changes attitudes: once informed, 58% of people said they are uncomfortable with how their conversations are used to train AI, with 30% describing themselves as “very uncomfortable,” according to the survey.

The data also shows a sharp divide between parents and non-parents. Parents or guardians with children in the household confide in AI at nearly twice the rate of non-parents. DuckDuckGo reports that “among parents who use AI, 43% have told an AI something they’ve never shared with another person,” while “that number drops to 25% for AI users with no children at home.”

DuckDuckGo places part of the responsibility on the AI companies themselves, arguing that they have leaned into confidant behavior through advertising campaigns and media appearances that encourage users to treat chatbots as trusted listeners, even when those conversations are less private than users assume. The company frames the survey as evidence of a mismatch between the intimacy people bring to these tools and the data-handling practices behind them.

Trust in the institutions holding that data runs low across the board. The survey found people have slightly more confidence in large AI companies than in the U.S. government when it comes to protecting their data, but skepticism dominates both. According to DuckDuckGo, 39% said they have no trust at all in large AI companies, compared with 48% who said the same of the government.

DuckDuckGo closes on the argument that AI privacy is still winnable ground precisely because chatbot habits remain new. The company draws a distinction between the questions already asked and the ones still to come. “The most valuable information you could share with Big Tech is your next question, not the questions you asked days or months or years ago,” DuckDuckGo states, adding that “it’s never too late to make the switch to more private services online.”

The company recommends that users take proactive steps to protect themselves, including choosing services with clearer and more transparent data-handling policies. The full findings are available through DuckDuckGo directly. For readers tracking how AI tools handle personal data, the survey underscores a widening gap between everyday use and everyday understanding, and it is a gap we are keeping an eye on as chatbot adoption continues to grow.