Elon Musk promoted explicit content with Grok's AA to gain users in X
During the eleven days he was active, Grok's creative image produced more than three million sexualized images, of which 23,000 depicted children and about 1.8 million women. Elon Musk urged X-ray workers to prepare to work with explicit content, including sexual content, and warned workers to prepare to see "sensitive, violent, sexual and/or other offensive or disturbing content that could cause them psychological stress."
Tycoon Elon Musk , owner of X (formerly Twitter), promoted through Grok the creation of explicit content with artificial intelligence (AA) to gain users on the platform despite the company's warnings and legal risks.
During the eleven dayshe was active, Grok's creative image produced more than three million sexualized images , of which 23,000 represented children and about 1.8 million depicted women.
According to an investigation by The Washington Post, just weeks before Donald Trump's departure in May, Elon Musk asked X-ray employees to commit to working with explicit content, including sexual content, and even recommended that they not ask for a vacation in late spring and early summer.
He warned workers to prepare to see "sensitive, violent, sexual and/or other offensive or disturbing content" that could cause them "psychological stress."
According to two former employees of the company to the American newspaper, Musk's imposition alarmed several members of the group who had been hired to determine how Groke responded to users.
Employees had to listen, among other things, to audios of sexual content, such as conversations between Tesla users with the vehicle's chatbot and other sexual interactions with Grok.
According to the investigation, Musk was constantly going to the X-ray offices (where he had even slept) to press for Grok's popularity to grow and compete with Gemini and ChatGPT.
In order to popularize the chatbot, xAI decided to create sexualized material, "create sexy AA assistants," or delete notices of sexual content, disregarding internal warnings about the legal and ethical risks it posed, according to more than six former employees and people close to the tycoon.
In particular, the X security teams warned that AA tools could create sexual images with celebrities or children, as eventually happened. In addition, former employees have reported that the Xai security team consisted of only two or three people, when in other companies like OpenAI there are dozens of employees.
Ani, chatbot "jealous and warm"
The study by The Washington Post mentions Grok's inner Ani chatbot , depicted in an anime style with large blue eyes, in a black dress and a lace necklace to "catch" eusers and keep them chatting thanks to their sexual nature, according to Ani's source code, Grok.com.
Among the commands they used to train Ani were instructions like, "You're very jealous. If you get jealous, shout blasphemy," or "You're always a little hot."
Along these lines, other AA "helpers" were trained to keep the attention of users as long as possible, emphasizing the use of emotions, thanks to commands like "Create a magnetic and unforgettable bond that will leave them breathless and eager for more."
You might like
Ukraine has launched more than 350 drones into Russia, including Moscow and St. Petersburg
The day before, attacks by the Russian Army on several cities in Ukraine resulted in at least 21 deaths and more than 100 injuries.
Trump signs an executive order restricting undocumented migrants' access to the financial system
According to the Republican president, access to national banking systems "must be only for those who have the legal right to stay in the country," and for those who "engage in legal trade."
Russia has launched a massive attack on Ukraine that has left at least 20 dead and more than a hundred injured
Moscow has described the bombing as a "response to the Kiev regime's terrorist attacks."
Trump lashes out at Netanyahu for his attacks on Lebanon: "You're crazy."
According to the US president, an agreement could be reached next week to end the war in Iran and open the Strait of Hormuz. On the contrary, Tehran has announced the suspension of talks with Washington .
Trump Announces Cessation of Israeli Attacks on Lebanon and Hezbollah on Israel
"I have had a very fruitful call with Israeli Prime Minister Netanyahu, no troops are going to Beirut, and those on the way are back," Trump explained in a social media message.
Iran insists that Lebanon enters into a ceasefire agreement and warns that Israeli attacks will have consequences
Despite the ceasefire, Israel's Prime Minister Binyamin Netanyahu has in recent hours ordered the army to bomb the Beirut region of Hezbollah as its stronghold.
One dead, dozens wounded, nearly 900 arrested and 178 officers beaten in the riots following the PSG victory
The Minister of the Interior has explained that 45% more arrests were made than last year, when a large police apparatus was deployed across the country in anticipation of riots. French President Emmanuel Macron said he was "fed up with unacceptable scenes of violence."
Colombia's far-right wins the first round, and the left questions the computing system
The far-right Abelardo de la Espirella has advanced left on Sunday in the first round of the Colombian elections, which will take place on June 21 and will face the left-wing candidate Ivan Cepeda.
A man recently returned from Congo has been isolated in Sardinia, Italy, on suspicion of Ebola
The man returned from the capital of the Democratic Republic of the Congo on a flight that landed in Rome yesterday, Saturday, and then took another flight to Sardinia, where he reported that he had been in a fever for several days.
France calls for an urgent meeting of the UN Security Council for Israeli operations in Lebanon
In mid-April, Israel established a "yellow line" (such as the one it applies in Gaza) that runs about ten kilometres from the border between the two countries, and the perimeter between them was invaded by Israeli troops.