San Francisco-based artist Olivia Tai launched a year-long participatory art project titled “What We Tell AI,” which has gathered over 350 anonymous handwritten confessions from the public about their interactions with artificial intelligence. Tai began the initiative in September 2025 after setting up a makeshift confession booth in Mission Dolores Park, where she invited strangers to share their private experiences with AI technology.
The project draws inspiration from the 2000s-era community mail art phenomenon PostSecret. Tai, who facilitates workshops on human-AI relationships, adapted the concept of anonymous disclosure for the digital age. Her booth appears in diverse locations across the Bay Area, New York City, and Denver, including parks, shopping malls, and AI-focused conferences.
According to a June 2026 Pew Research Center survey, approximately half of U.S. adults now report using AI chatbots such as OpenAI’s ChatGPT, Google’s Gemini, or Microsoft’s Copilot, a significant rise from one-third of adults in mid-2024. Roughly 25% of American adults use these tools daily.
Tai describes the collection as an “ethnographic snapshot” rather than a comprehensive representation of public opinion. The submissions reveal a complex landscape of user behavior, ranging from pride to shame. Common themes include using AI for dating advice, coping with breakups, and seeking emotional support. Some contributors reported forming romantic attachments to AI entities, while others expressed guilt about relying on chatbots for personal issues.
One notable submission described a user cloning an ex-partner’s voice to achieve closure, while another from Palo Alto read, “Forgive my sin: I talk to AI way more than to ALL people in my life combined.” Academic research supports these observations; an AI Observatory analysis of over 24,000 conversations found that users increasingly turn to chatbots for health and relationship guidance.
Tai worries that the privatization of emotional discourse through AI may lead to a gradual erosion of human social networks. She cites data from Elon University and The Washington Post indicating that 27% of Americans use chatbots for personal and emotional topics, with nearly 40% of those users stating they share information with AI that they would not tell other people.
Despite concerns about the societal impact, Tai’s goal is to normalize discussions about AI usage. She argues that the current binary discourse—ranging from uncritical adoption to neo-Luddite rejection—marginalizes those in the middle who feel ashamed of their reliance on the technology. By removing stigma, Tai hopes individuals can feel more empowered to engage with broader ethical questions regarding AI training, data centers, and regulation.
Leave a Reply