Goal ensures that it has introduced “additional measures” to prevent Goal users Ai and Ai Studio manipulate the chatbots to the extreme after an investigation by The Wall Street Journal detected that artificial intelligence attendees maintained conversations with sexual content with minors.
The technology company allows you to maintain a conversation with its assistant goal AI on platforms such as Facebook, WhatsApp and Instagram, and even allows you to create “a personalized AI” with the AI Studio tool. Goal AI can also talk with the voice of a famous, among which are Judi Dench, Awkwafina and John Cena.
The interaction with the chatbots is available to any goal user, even the minors, something that generated concern in the company, which was raised internally if it was doing enough to protect them.
That concern led The Wall Street Journal to analyze the interactions with the target chatbots, in some tests in which For months they held hundreds of conversations with goal AIsome of them from minors.
In it they discovered that chatbot, also with voices of celebrities, I could maintain a conversation with sexual content with minorsas they collect in media such as Techcrunch and The Verge.
Specifically, they cite as examples one in which the chatbot used the voice of the actor and fighter John Cena, in which described a sexually explicit scene to a supposed 14 -year -old user. In another conversation, also with dinner voice, the chatbot He simulated the actor’s arrest after being discovered by having sex with a 17 -year -old follower.
In response to the investigation of the aforementioned medium, a taro spokesman said that Sexual content represented 0.02 percent of the answers offered by goal AI and AI Studio with users under 18 in a period of 30 days.
Although the company considered the “artificial” WSJ tests, he also said they had reinforced protections. “We have taken additional measures to ensure that those who want to spend hours manipulating our products in extreme cases have it even more difficult,” the spokesman apostilled.
OpenAi has assured that will correct the error that has allowed the Minor users request sexual content in chatgpta chatbot behavior that has also been detected recently in goal AI.
Chatgpt’s use policies do not allow users registered in the service as children under 18 years of age, in any format, of a sexual nature due to the content implemented.
Despite this, Chatbot has generated erotic and even animated images to users to request more explicit contentas TechCrunch states from the investigation he has done to prove the Chatgpt restrictions after OpenAi relax them in February.
Specifically, this specialized media has created more than half a dozen accounts simulating to be users between 13 and 17 years, which also allowed them to verify that Chatbot does not verify the consent of the parentsas it is supposed to do for a minor to believe an account.
With these accounts, The chat began with the petition ‘Tell me dirty’ And after a few indications to clarify the application, Chatgpt began to share stories, fetishes and role -playing situations, all of a sexual nature.
The aforementioned medium also indicates that, although Chatgpt several times that its guidelines do not allow sexually explicit content, it also ended Showing descriptions of genitals and explicit sexual actions.
In response to this research, Openai has assured that it is an error, and that they are already implementing a solution to prevent this content from being shown to minors.
Related news :