The UK’s info watchdog will examine reviews that Elon Musk’s AI chatbot, Grok, has been used to generate sexual imagery of youngsters.
Grok was developed by Musk’s xAI in 2023, designed to be a “truth-seeking” assistant with a witty, rebellious persona.
Built-in into X, previously Twitter, it makes use of real-time information from the platform to generate textual content, photographs and code.
However complaints have mounted that Grok was being used to generate sexual photos of actual ladies and youngsters, and now the Info Commissioner’s Workplace (ICO) is investigating.
The announcement comes on the identical day the X offices in Paris were raided by French prosecutors inspecting related allegations.
In an announcement on its web site, the ICO confirmed it had opened a proper probe into two X firms regarding their processing of non-public information in relation to Grok, and the AI’s potential to supply dangerous sexualised picture and video content material.
“We now have taken this step following reviews that Grok has been used to generate non‑consensual sexual imagery of people, together with kids,” the assertion mentioned.
“The reported creation and circulation of such content material raises severe considerations underneath UK information safety legislation and presents a danger of great potential hurt to the general public.”
File pic: Reuters
William Malcolm of the ICO mentioned the investigation would probe whether or not X Web Limitless Firm and xAI had complied with information safety legal guidelines, and offered adequate safeguards.
He mentioned: “The reviews about Grok increase deeply troubling questions on how individuals’s private information has been used to generate intimate or sexualised photographs with out their information or consent, and whether or not the required safeguards have been put in place to forestall this.
“Shedding management of non-public information on this approach may cause quick and vital hurt. That is significantly the case the place kids are concerned.
“Our function is to deal with the info safety considerations on the centre of this, whereas recognising that different organisations even have essential obligations.”
Ofcom additionally investigating
One other regulatory watchdog, Ofcom, additionally opened a proper investigation into X final month underneath the UK’s On-line Security Act, to find out whether or not the agency was complying with its duties to guard individuals from unlawful content material.
The European Fee launched a probe into Grok final month too, taking a look at whether or not it disseminates unlawful content material, resembling manipulated sexualised photographs, within the EU.
The EU is one in every of a variety of authorities around the globe to have raised considerations about Grok, with authorities officers in Germany, Sweden, India, Japan, Malaysia, California, Indonesia and Philippines amongst those who have spoken out.
Mr Malcolm mentioned the ICO was working carefully with Ofcom and “worldwide regulators”.
File pic: iStock
Grok has restricted picture modifying however but to say the place
In a separate assertion, regulator Ofcom mentioned – whereas it was probing X – it was not investigating xAI, which supplies the standalone Grok chatbot app.
Ofcom additionally mentioned its investigation into X was nonetheless gathering proof and warned the probe may take months.
On why it was not investigating xAI, the assertion mentioned: “After we opened our investigation into X, we mentioned we have been assessing whether or not we also needs to examine xAI, because the supplier of the standalone Grok service.
“We proceed to demand solutions from xAI in regards to the dangers it poses. We’re inspecting whether or not to launch an investigation into its compliance with the foundations.”
Ofcom mentioned that due to the way in which the On-line Security Act pertains to chatbots, it was at present unable to research the creation of unlawful photographs by the standalone Grok itself.
xAI mentioned on 14 January it had restricted picture modifying for Grok AI customers and blocked customers, primarily based on their location, from producing photographs of individuals in revealing clothes in “jurisdictions the place it is unlawful”. It’s but to determine the international locations the place these restrictions apply.
xAI earlier mentioned it had restricted the usage of Grok’s picture technology and modifying options to paying subscribers solely.
