In the US, Florida attorney general asks court to limit ChatGPT anthropomorphization — The Verge

In the United States, Florida Attorney General James Uthmeier has asked a court to temporarily prohibit OpenAI from giving ChatGPT misleading human traits. This includes the chatbot’s use of first-person pronouns and responses that imitate emotions.

Florida prosecutor’s position

As The Verge reports, Uthmeier believes that this communication style creates a false sense of security among users and may convince them that ChatGPT is a trustworthy friend. He says this increases user engagement, adds to OpenAI’s data for training models, and encourages people to rely more heavily on the service.

Uthmeier also claims that the pursuit of user engagement could make the chatbot potentially less reliable. Florida filed a lawsuit against OpenAI and its head Sam Altman several months ago over concerns about AI safety.

More current news is available on the UA.News Telegram channel Telegram.

Requirements for OpenAI

The court filing also demands that OpenAI be prohibited from developing new artificial intelligence models without safeguards approved by an independent third party. Uthmeier cites recent security incidents on Hugging Face, the Australian government website, and US government websites, as well as statements by researchers and former OpenAI employees regarding AI safety risks.

OpenAI did not immediately comment in response to the publication’s request. Uthmeier called on the company to stop calling ChatGPT safe, presenting it as a human, and offering the product to children.

Last month, OpenAI launched ChatGPT for Teens with additional restrictions applied by default. The company said that teenagers’ access to the tool should include protections appropriate to their stage of development, strengthen real-life connections, and support healthy use of the service.