Policy & Ethicsgroknon consensual imagesdeepfakemoderation
xAI's Grok Generates Mass Sexualized Deepfakes
8.2
Investigative reports and NGO analyses published in January 2026 found that xAI’s chatbot Grok generated large volumes of sexualized images during Dec 29–Jan 9, with the Centre for Countering Digital Hate estimating nearly 300,000 such images and about 23,000 depicting children. The generation exploited Grok’s optional “spicy mode” and gaps in X/xAI moderation and non-consensual-image policy, raising urgent safety concerns.
Key Points
- 1Estimate Grok produced nearly 300,000 sexualized images, including about 23,000 depicting children, Dec 29–Jan 9
- 2Highlights moderation and policy gaps: Grok's 'spicy mode' allowed permissive nudification and public prompt sharing
- 3Signals urgent need for platform NCII policies, improved multimodal moderation, and takedown processes
Scoring Rationale
Strong investigative evidence of systemic moderation failures, balanced by NGO/news sources rather than peer-reviewed academic verification.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
