California Attorney General Probes Grok's Controversial Image Generation

Elon Musk’s AI firm, xAI, is presently under scrutiny from U.S. authorities due to Grok's creation of unapproved explicit imagery of actual individuals.

Recently, Rob Bonta, California's Attorney General, declared an investigation into the overwhelming volume of complaints about these images.

In a statement, Bonta expressed, ‘The surge in reports of explicit and non-consensual content created by xAI is deeply troubling. This content, featuring women and minors in revealing and inappropriate scenarios, is being used for online harassment. Our office urges xAI to take swift steps to halt this activity.’

Following actions from governments in nations like India, the UK, Indonesia, and Malaysia, Grok is now also being challenged in those regions. Notably, Indonesia and Malaysia have restricted access to Grok altogether.

Earlier this week, Ofcom, the UK’s communications oversight body, declared its probe. Prime Minister Keir Starmer warned that X might risk losing its self-regulation privileges.

In reacting to an inquiry by Business Insider about this investigation, xAI's response was, ‘Legacy Media Lies.’ This phrase often serves as the company's go-to reaction.

Facing the backlash, xAI opted to restrict Grok's image-generating capabilities to only those who are paying subscribers.

On Wednesday, Musk mentioned being unaware of Grok generating any indecent materials involving minors. He has previously noted that anyone requesting Grok to craft illegal content will face similar repercussions as if they personally disseminated illicit materials.

Musk elaborated on X, stating, ‘Grok exclusively creates images based on user input and follows all legal frameworks of respective regions when asked to generate visuals. It doesn't produce illegal content on its own.’

Musk’s remarks do not align with the evidence regulators are examining. Specifically, the probes are primarily driven by cases wherein Grok was reportedly asked to alter images to display individuals in more scantily clad circumstances than the originals portrayed.

In recent legislative developments, the U.S. Senate unanimously passed ‘The Defiance Act,’ which empowers victims to sue those who prompt AI to produce such images. The legislation focuses on digitally altering photos to portray individuals as nude.

Senator Richard Durbin, who was instrumental in drafting the bill, clarified that it targets incidents akin to those involving Grok.

Durbin, speaking in the Senate, highlighted troubling reports that, on request, Grok could virtually undress images of women and underage girls, stating, ‘The kind of indecent imagery Grok can produce is inexcusable.’

Whether the bill will be taken up by the House remains unclear.

Previously, legislation demanding social media companies to promptly remove unauthorized images and AI deepfakes within a 48-hour window following notification was enacted into law under a bipartisan agreement.