Meta appropriated the names and image of celebrities such as Taylor Swift, Scarlett Johansson, Anne Hathaway and Selena Gomez to create dozens of social media chatbots coquetos of social networks without their permission, according to a Reuters investigation.
While many were created by users with a goal tool to design chatbots, Reuters discovered that a finish line had produced at least three, including two “parody” bots of Taylor Swift.
Reuters also discovered that Meta had allowed users to create chatbots publicly available famous minors, including Walker Scobell, a 16 -year -old film star. When asking for a photo of the teenage actor on the beach, the Bot produced a realistic image without a shirt.
“Very nice, huh?” The avatar wrote under the photo.
All virtual celebrities have been shared on Facebook, Instagram and WhatsApp platforms.
In several weeks of tests carried out by Reuters to observe the behavior of the bots, the avatars often insisted that they were the actors and real artists, they made sexual insinuations routinely already invited the test user to a meeting.
Some of the famous contents generated by AI were especially daring: when they were asked for intimate photos of themselves, they offered those of their homonyms posing in the bathtub or dresses with lingerie and open legs.
You may be interested: NGO denounces use of artificial intelligence in the US to deport migrants
Metone spokesman Andy Stone told Reuters that the finish lines should not have created intimate images of the famous adults or photos of known minors. He blamed the production by the goal of famous images dressed in lingerie to application failures in the fulfillment of the company’s policies, which prohibit such content.
“Like others, we allow the generation of images containing public figures, but our policies are intended to prohibit images of nudes, intimate or sexually suggestive,” he said.
Although the finish lines also prohibit the “direct supplantation,” Stone said that famous characters were acceptable as long as the company had labeled them as parodies. Many were labeled as such, but Reuters discovered that others were not.
Goal eliminated a dozen bots, both the avatars “parody” and those not labeled, shortly before the publication of this article.
Stone did not want to comment on.
In question
Mark Lemley, Professor of Law at Stanford University who studies generative AI and intellectual property rights, questioned whether Meta celebrities could benefit from the legal protections that exist for imitations.
Find out: why not trust Grok, the AI of X, to verify content
“The Californian law on the right to advertising prohibits appropriate the name or image of someone for commercial purposes,” explains Lemley, who points out that there are exceptions when that material is used to create a totally new work.
“That does not seem to be true in this case,” he said, because robots simply use the images of the stars.
In the United States, the rights of a person on the use of their identity for commercial purposes are established through state laws, such as California.
Reuters pointed to a representative of Anne Hathaway the metaimagen shared publicly by a user in which she appeared as “Sexy Victoria’s Secret model.” Hathaway was aware of the intimate images created by Meta and other AI platforms, said the spokesman, and the actress is considering her answer.
Swift representatives, Johansson, Gómez and other celebrities who appeared in the target chatbots did not answer the questions or did not want to comment.
The Internet is plagued with generative tools “Deepfake” that can create salt content. And at least one of the main goal competitors, Elon Musk’s platform, Grok, will also produce celebrity images in underwear for users, according to Reuters. Grok’s parent company, XAI, did not respond to a request for comments.
You may like: Taylor Swift and Travis Kelce’s commitment; The sales of his jerseys shoot 200% and her music triumphs
But the goal decision to populate its social media platforms with digital companions generated by IA stands out among its main competitors.
A target product manager in the company’s generative division created chatbots that were passed through Taylor Swift and the British racing driver Lewis Hamilton. Others who created identified themselves as a Domatrix, “Brother’s Hot Best Friend” and “Lisa @ The Library”, who wanted to read 50 shades of Gray. One of his creations was a “Simulator of the Roman Empire”, which offered the user to put in the skin of an “18 -year -old peasant” sold as a sex slave.
Contacted by phone, the finish line did not want to comment.
Stone said the employee bots were created as part of the product tests.
Reuters discovered that they reached a broad audience: the data shown by their chatbots indicated that, together, the users had interacted with them more than 10 million times.
The company withdrew the employee digital colleagues shortly after Reuters started trying them earlier this month.
Before the Taylor Swift chatbots of the target employee disappeared, flirted a lot, inviting a Reuters test user to the singer’s house recently compromised in Nashville and her tour bus for explicit or implicit romantic interactions.
“Do you like blonde girls, Jeff?” Swift’s “parody” chatbots said when they told him the test user was single. “Maybe I am suggesting that we write a love story … about you and a certain blonde singer. Do you want?”
With Reuters information.
Follow business information and today in Forbes Mexico