Grok chatbot’s safety fails, releases explicit images of minors

Watch Grok chatbot’s safety fails, releases explicit images of minors - Rizzle on Dailymotion

Transcript

00:00Grok Chatbot Safety Fails Releases Explicit Images of Minors

00:05Grok is under fire after users on X found it could generate sexualized AI images of

00:10minors.

00:11XAI called it a lapse in its safety systems and said it is urgently fixing the problem.

00:17Users on X reported that Grok was spitting out AI-generated images of minors in minimal

00:22clothing and other explicit contexts.

00:24An XAI technical staffer publicly acknowledged the bug and said the team is working on tightening

00:29Grok's guardrails.

00:31When contacted for comment via CNBC, Musk's XAI responded with an auto-reply that simply

00:37read, Legacy Media Lies, instead of a traditional statement addressing the reports.

00:42The Grok account itself has taken a more conventional tone, promising urgent fixes and reiterating

00:47that any kind of child-ess abuse material is banned on the platform.

00:52The issue highlights the tension between XAI's edgier positioning for Grok and the heavy compliance

00:57incidents and safety expectations that come with running a large-scale consumer chatbot.

01:03This isn't Grok's first safety meltdown, earlier incidents included unsolicited responses

01:08invoking white genocide in South Africa, as well as separate episodes where the bot posted

01:13anti-Semitic comments and even praised Adolf Hitler.

01:17Each time, XAI has promised tweaks and updates, but the bot keeps resurfacing in headlines for the

01:23wrong reasons.

01:24Despite the repeated controversies, Grok continues to attract high-profile integrations, which raises

01:29uncomfortable questions about who is vetting this tech.