Scott Wiener Removes AI Chatbot Mimicking Opponent Connie Chan

California congressional candidate Scott Wiener removed an AI chatbot that mimicked the voice and accent of his opponent, Connie Chan, this month [1, 2].

The incident highlights growing concerns over the use of generative artificial intelligence in political campaigns to create deceptive or harmful content. As candidates seek new ways to reach voters, the line between political parody and targeted harassment has become a central point of contention in U.S. elections.

The chatbot was deployed in early August [2, 3]. According to Wiener, the tool was designed to highlight the record of Chan on various policy issues, including public housing [1, 2]. However, the campaign faced immediate backlash from critics who said the bot did not focus on policy but instead mocked Chan's accent [1, 2].

Those criticizing the bot said the mimicry carried racist and sexist connotations [1, 2]. The controversy unfolded in the San Francisco congressional district, a seat formerly held by Nancy Pelosi [2, 4]. The backlash led to a rapid reversal by the Wiener campaign, which deleted the AI content after it had been active for about one week [2, 3].

While the campaign framed the bot as a parody, the reaction from observers suggested a failure to account for the cultural implications of voice synthesis. The removal of the bot follows a period of intense scrutiny over how AI can be used to impersonate political figures, especially those from minority backgrounds, without their consent [1, 2].

Wiener said the intent was to discuss political records, but the execution resulted in a tool that critics viewed as a caricature [1, 2]. The event marks one of the first high-profile instances in this specific race where AI-generated persona mimicry led to a formal campaign retreat [2, 3].