Technology
DuckDuckGo survey shows many users share highly personal data with AI chatbots
A survey of nearly 2,000 U.S. adults reveals a widespread lack of awareness about how AI companies handle and store user conversations.
The short version
- A survey by search engine DuckDuckGo found that nearly one-third of AI chatbot users have shared information with AI that they would not disclose to friends, family, colleagues, or doctors.
- More than half of the respondents were unaware or unsure that most AI chatbots use their conversations for training purposes by default.
- A significant majority of respondents reported feeling uncomfortable after learning that their chat histories are utilized to train AI models.
Key facts
- DuckDuckGo surveyed nearly 2,000 U.S. adults to examine privacy awareness and habits regarding AI chatbots.[9to5Mac]
- Among self-described AI enthusiasts, the rate of sharing sensitive secrets with AI that they would not tell friends, family, colleagues, or medical professionals rises to 56%.[9to5Mac]
- Forty-three percent of AI-using parents or guardians with children at home reported confiding something to an AI that they have never shared with another person, compared to 25% of AI users without children at home.[9to5Mac]
- Only 25% of survey respondents knew that AI chatbot conversations can be subpoenaed by law enforcement.[9to5Mac]
- Fifty-eight percent of respondents expressed discomfort when informed that their conversations are used to train AI.[9to5Mac]
- The survey indicated high levels of skepticism toward data protection, with 39% of respondents expressing zero trust in large AI companies and 48% expressing zero trust in the U.S. government.[9to5Mac]