UNICEF warns of a surge in AI-generated sexual deepfakes involving children, including nudification that removes or alters victims' clothing.

Policymakers are urged to broaden legal definitions to cover AI-generated child sexual abuse content and criminalize its creation, acquisition, possession, and distribution.

The agency calls for safety-by-design approaches and guardrails from AI developers, plus stronger content moderation and detection tech from platforms to curb circulation.

The piece is framed as a news brief with promotional content for services, not core article content.

Reporting credits go to Jasper Ward in Washington, with editorial work by Michelle Nichols and Rod Nickel.

xAI has restricted Grok AI image editing in jurisdictions where it’s illegal and limited certain features to paying subscribers as part of a misuse-curb effort.

Grok’s image generation and editing features are limited to paying subscribers and blocked in places where such activity is illegal.

Country examples studied include Armenia, Brazil, Colombia, Dominican Republic, Morocco, Montenegro, North Macedonia, Mexico, Pakistan, Serbia, and Tunisia.

While some developers adopt safety-by-design, overall safeguards are patchy and risks rise when generative AI tools are embedded in social media.

UNICEF and others note progress on safeguards is uneven, with higher risks when AI tools spread via social platforms.

UN Secretary-General proposed an Independent International Scientific Panel on AI to guide governance, emphasizing guardrails, innovation for the common good, and global cooperation.

Britain unveiled plans to criminalize using AI tools to create child sexual abuse images, signaling a potential global regulatory precedent.