A ChatGPT-based bot imitated a US presidential candidate: OpenAI has taken action
OpenAI has banned the developer of an artificial intelligence bot that imitates US Democratic presidential candidate Dean Phillips (pictured above). This is the first response from ChatGPT to what it considers to be the misuse of its AI tools in a political campaign.
The Dean.Bot chatbot was created by entrepreneurs Matt Krisiloff and Jed Somers in support of Phillips ahead of the New Hampshire primary election, with financial support from hedge fund manager Bill Ackman, an activist and billionaire. The bot was based on the GPT language model from OpenAI.
The authors of the idea signed a contract with Delphi, an artificial intelligence startup, to create a bot. OpenAI suspended Delphi’s account and noted that OpenAI rules prohibit the use of its technology in political campaigns. Delphi liquidated Dean.Bot after the account was suspended.
“Anyone who creates with our tools must adhere to our usage policy. We recently deleted a developer account that knowingly violated our API terms of use, which prohibit political campaigning or impersonation without consent,” says OpenAI spokesperson Lindsey Held.
The bot had a warning that it was an artificial intelligence tool, not the real Dean Phillips, and demanded that the voter agree to its use. But such technologies can make people mistake a tool like this for a person, even if there are reservations.
Proponents of this method of campaigning argue that bots, if used properly, can educate voters by providing them with an interesting way to learn more about a candidate.
According to experts, such technologies can provoke a wave of automated calls to voters who think they are talking to real candidates. Artificial intelligence systems can also produce disinformation on a large scale.
Customers asked Delphi to remove ChatGPT from the bot and instead rely on open source technologies that also offer conversational capabilities.
