WASHINGTON: Elon Musk’s AI software Grok confronted rising worldwide backlash on Monday (Jan 5) for producing sexualised deepfakes of ladies and minors, with the European Union becoming a member of the condemnation and Britain warning of an investigation.
Complaints of abuse flooded the web after the current rollout of an “edit picture” button on Grok, which enabled customers to change on-line photographs with prompts comparable to “put her in a bikini” or “take away her garments”.
The digital undressing spree, which follows rising considerations amongst tech campaigners over proliferating synthetic intelligence (AI) “nudify” apps, prompted swift probes or requires remedial motion from international locations together with France, India and Malaysia.
The European Fee, which acts because the EU’s digital watchdog, joined the refrain on Monday, saying it was “very critically trying” into the complaints about Grok, developed by Musk’s startup xAI and built-in into his social media platform X.
“Grok is now providing a ‘spicy mode’ exhibiting specific sexual content material with some output generated with childlike photographs. This isn’t spicy. That is unlawful. That is appalling,” stated EU digital affairs spokesman Thomas Regnier.
“This has no place in Europe.”
The UK’s media regulator Ofcom stated it had made “pressing contact with X and xAI to grasp what steps they’ve taken to adjust to their authorized duties to guard customers within the UK”.
Relying on the reply, Ofcom will then “decide whether or not there are potential compliance points that warrant investigation”.
“HORRIFYING”
Malaysia-based lawyer Azira Aziz expressed horror after a consumer – apparently within the Philippines – prompted Grok to alter her “profile image to a bikini”.
“Harmless and playful use of AI like placing on sun shades on public figures is okay,” Aziz informed AFP.
“However gender-based violence weaponising AI towards non-consenting ladies and kids have to be firmly opposed,” she added, calling on customers to report violations to X and Malaysian authorities.
Different X customers immediately implored Musk to take motion towards obvious paedophiles “asking grok to place bikinis on kids”.
“Grok is now undressing pictures of me as a toddler,” Ashley St Clair, the mom of certainly one of Musk’s kids, wrote on X.
“That is objectively horrifying, unlawful.”
When reached by AFP for remark, xAI replied with a terse, automated response: “Legacy Media Lies.”
Amid the net firestorm, Grok sought to guarantee customers on Friday that it was scrambling to repair flaws within the software.
“We have recognized lapses in safeguards and are urgently fixing them,” Grok stated on X.
“CSAM (Baby Sexual Abuse Materials) is prohibited and prohibited.”
Individually final week, Grok posted an apology for producing and sharing “an AI picture of two younger ladies (estimated ages 12-16) in sexualised apparel primarily based on a consumer’s immediate”.
“GROSSLY OFFENSIVE”
The flurry of reactions got here after the general public prosecutor’s workplace in Paris final week expanded an investigation into X to incorporate new accusations that Grok was getting used for producing and disseminating youngster pornography.
The preliminary investigation towards X was opened in July following reviews that the platform’s algorithm was being manipulated for the aim of overseas interference.
On Friday, Indian authorities directed X to take away the sexualised content material, clamp down on offending customers and submit an “Motion Taken Report” inside 72 hours, or face authorized penalties, native media reported.
The deadline lapsed on Monday, however thus far there was no replace on whether or not X responded.
The Malaysian Communications and Multimedia Fee additionally voiced “severe concern” on the weekend over public complaints concerning the “indecent, grossly offensive” content material throughout X.
It added it was investigating the violations and can summon X’s representatives.
The criticism provides to rising scrutiny of Grok, which has confronted criticism for churning out misinformation about current crises such because the battle in Gaza, the India-Pakistan battle and a lethal taking pictures in Australia.
