OpenAI bans chatbot impersonating U.S. presidential candidate’s voice
OpenAI bans a chatbot impersonating a US presidential candidate’s voice, following strict limits on its AI capabilities
OpenAI, the creator of ChatGPT, has removed a chatbot impersonating a lesser-known US presidential candidate’s voice. The move comes as OpenAI enforces strict limits on the political use of its AI capabilities.
A political group supporting US Congressman Dean Phillips had commissioned a Miami-based startup, Delphi, to create a chatbot allowing users to engage in a Q&A with the candidate. OpenAI confirmed the removal of a developer account violating its policies, which prohibit political campaigning and impersonation without consent.
“We recently removed a developer account that was knowingly violating our API usage policies,” OpenAI stated to the Washington Post, which broke the story. The “Dean-bot” was part of Phillips’ Democratic primary campaign against President Joe Biden.
Concerns over the misuse of generative AI, including ChatGPT, for political disinformation have grown, especially with upcoming major elections globally. OpenAI, as a leader in the field, faces heightened scrutiny and announced plans last week to provide users with tools to mitigate potential harm.
The startup Delphi, known for creating bots mimicking personalities like influencers or celebrities, relied on OpenAI’s technology. OpenAI promptly disabled the bot after the Washington Post shared a link to test it.
The political group behind the chatbot, “We Deserve Better,” co-led by a former OpenAI CEO Sam Altman staffer, has Silicon Valley investor backing, according to the Post.
In response to the incident, Delphi announced it would prohibit political organizations or individuals from using its platform. The startup acknowledged its mistake, stating, “We believed, incorrectly, that it would be acceptable under the OpenAI terms of service to have a political action committee that supports Dean Phillips create a clone of him using our platform.”
Delphi has issued apologies to both OpenAI and “We Deserve Better” for the error. After the controversy, Delphi was barred from accessing OpenAI’s platform. OpenAI continues to face challenges in balancing the powerful capabilities of its technology with responsible use, especially in politically sensitive contexts.
