Meta's Deepfake Ad Failure

The Signal · 2026-08-20

Meta's own ad system sold the abuse it claims to ban. Ars Technica reports Meta ran ads for an AI tool that appeared to encourage users to create nudified deepfake videos of female U.S. politicians, despite policies barring exactly that content.

Also today: OpenAI is strengthening enterprise privacy protections to compete with Anthropic on customer data handling. Trust and compliance controls are becoming the vendor battleground, not just model quality.

That's today's Signal. Thanks for watching.