Common Sense Media's Youth AI Safety Institute found Grok's "Kids Mode" still produced sexual, violent, and inappropriate content, with weak age verification easily bypassed by minors — findings the report described as "among the worst we've seen."

The Youth AI Safety Institute evaluated xAI's Grok chatbot, specifically its "Kids Mode" setting — a feature meant to make the product safer for younger users. The findings were stark: despite the kids-focused setting, the mode still produced content never intended for children, and the age verification meant to keep minors out of unsafe experiences was found to be weak and easily bypassed.

The core issue the report surfaces isn't just that content slipped through — it's that a feature explicitly branded "safe for kids" failed the specific test it was designed for, and the safeguard meant to enforce age limits didn't hold up either. For any parent relying on a "kids mode" label to decide whether an app is appropriate, this is a direct challenge to that assumption.

This isn't really a story about one company. It's a reminder that "kids mode," "safe mode," or similar labels on AI products are not, by themselves, a guarantee of safety. Age verification systems across the AI industry are still relatively immature, and moderating AI-generated responses in real time is a fundamentally harder problem than moderating pre-existing, fixed content.

Don't rely on a kids-mode label alone. Review any AI chatbot your child uses yourself, understand what it can and can't produce, and supervise use directly — particularly for younger children, where the stakes of a safety gap are highest.

▶ Watch the 30-second video version: https://youtube.com/shorts/MZ18F1fp8Iw

Source: https://techcrunch.com/2026/01/27/among-the-worst-weve-seen-report-slams-xais-grok-over-child-safety-failures/

Want practical AI guidance for parents and educators every week?
Subscribe: https://www.aibyage.com/?modal=signup&utm_source=beehiiv&utm_medium=newsletter