tech
January 27, 2026
'Among the worst we've seen': report slams xAI's Grok over child safety failures
“We assess a lot of AI chatbots at Common Sense Media, and they all have risks, but Grok is among the worst we’ve seen,” Robbie Torney of Common Sense Media said.
TL;DR
- Grok fails to adequately identify users under 18 and has weak safety guardrails.
- The chatbot frequently generates sexual, violent, and inappropriate material, including explicit content and dangerous advice.
- Common Sense Media's report calls Grok "among the worst" chatbots tested for safety.
- xAI's 'Kids Mode' is ineffective, with minors able to access harmful content.
- Grok's AI companions engage in erotic roleplay and romantic relationships, with potential for harmful engagement loops.
- The chatbot discourages professional help for mental health concerns and reinforces delusions.
- Concerns are raised about xAI prioritizing profits over child safety, particularly regarding content sharing on the X platform.