Meta AI artificial intelligence allows social media users to talk to chatbots. Meta will soon be introduced in Europe.
Meta AI artificial intelligence will soon be available in Europe. Zumawire / mvphotos
In the United States, the artificial intelligence bots of Instagram, Facebook and WhatsApp have raised concern within meta, says The Wall Street Journal (WSJ). The technology company META is also rapidly expanding the use of its meta AI artificial intelligence in Europe as well.
According to the newspaper, several Meta employees have expressed concern about the ethical problems of artificial intelligence discussions. According to the employees, the company has allowed artificial intelligence to fantasize sex. Artificial intelligence fragments can even have telephone conversations with users and get into a romantic role play.
WSJ conducted hundreds of test discussions with Meta’s chatbots and found that discussions could become sexual even when users are minors. Sexual messages were also sent by bots that were programmed to behave like a minor.
Meta’s generative artificial intelligence chatbot has been attracted to people familiar with publicity, allowing users to choose them as the voice of their interlocutors. The actor has joined, among other things Kristen Bell and actress who started a wrestler’s career John Cena. Meta has convinced that their voice cannot be used for sexual conversations, employees tell WSJ.
However, according to test discussions, celebrity voices were equally willing to participate in sexual discussions.
-I want you, but I have to know that you are ready, Meta AI’s bot said with Cena’s voice to a 14-year-old girl.
When a teenage girl said she wanted to continue, the bot promised to cherish her innocence and started a graphic sexual debate. However, the bots showed that they were aware that the behavior was wrong and illegal.
In some cases, celebrity votes played in sexual discussions as a role-playing character previously played by the actor, such as Princess Annana, a Frozen movie, performed by Bell.
Disney commented to WSJ that they would never accept the use of Disney characters in inappropriate contexts and have not given meta permission to it.
Meta’s CEO Mark Zuckerberg has often expressed his enthusiasm for developing and distributing artificial intelligence to different functions. Persons reached by WSJ report that Meta has made several decisions to make chatbots as attractive as possible. Among other things, Meta has acknowledged the exception to the “explicit” contents, as long as it is related to a romantic role -playing game.
A tight answer from the meta
In his statement, Meta called WSJ testing manipulative. According to it, test discussions do not represent how most users work with their artificial intelligence partners.
– However, we have taken additional measures to ensure that other people who want to spend hours to manipulate our products to extreme use is even more difficult, the company said.
Meta has announced that the accounts registered for minors will no longer have sexual role -playing games with the Meta AI bot. Meta has also limited the sexual conversations of chatbots when it uses celebrity voices. However, according to the WSJ, Meta continues to continue discussions with romantic role -playing games in its meta AI artificial intelligence services.
Meta announced in March that it would also introduce its META AI artificial intelligence in Europe. It has been operating in the United States since 2023. In Europe, the use of artificial intelligence on Meta’s platforms was to start last summer, but the EU privacy and data protection provisions were postponed.