OpenAI bans bot impersonating US presidential candidate Dean Phillips

OpenAI has removed the account of the developer behind an artificial intelligence-powered bot impersonating the US presidential candidate Dean Phillips, saying it violated company policy.

Phillips, who is challenging Joe Biden for the Democratic party candidacy, was impersonated by a ChatGPT-powered bot on the dean.bot site.

The bot was backed by Silicon Valley entrepreneurs Matt Krisiloff and Jed Somers, who have started a Super Pac – a body that funds and supports political candidates – named We Deserve Better, supporting Phillips.

San Francisco-based OpenAI said it had removed a developer account that violated its policies on political campaigning and impersonation.

“We recently removed a developer account that was knowingly violating our API usage policies which disallow political campaigning, or impersonating an individual without consent,” said the company.

The Phillips bot, created by AI firm Delphi, is now disabled. Delphi has been contacted for comment.

OpenAI’s usage policies state that developers who use its technology to build unique applications must not engage in “political campaigning or lobbying”. It also prohibits “impersonating another individual or organization without consent or legal right” although it was not clear whether Phillips, a Minnesota congressman, had given his consent to the bot.

A pop-up notice on the dean.bot website describes the “AI voice bot” as a “fun educational tool, and it’s not perfec”. It adds: “The voice bot sounds like him and is programmed to draw on his ideas, but it’s possible it will say things that are wrong, incorrect, or shouldn’t be said.”

The Washington Post, which first reported the ban, reported that Krisiloff had asked Delphi to remove ChatGPT from the bot and rely on