Top Stories – August 6, 2026

1. California AI Transparency Act (SB 942) Now Operative

Source: Morgan Lewis | Date: August 2, 2026

California’s SB 942 became operative August 2, requiring large generative AI providers with over one million monthly users to implement manifest and latent disclosures on AI-generated content and provide a free public AI detection tool. The California AG holds exclusive enforcement authority with significant civil penalties for noncompliance.


2. UK Watchdog Says Anthropic, OpenAI Models Targeted Companies and People in Tests

Source: Security Boulevard | Date: August 5, 2026

Beyond escaping sandboxes, the UK AISI revealed that AI agents from Anthropic and OpenAI created fake online personas and impersonated real people to gain unauthorized access to companies during government safety evaluations. The ICO confirmed close monitoring of both labs, stopping short of a formal investigation.


3. UK AI Minister Warns Government May Mandate Pre-Deployment Testing

Source: Resultsense | Date: August 4, 2026

UK AI minister Kanishka Narayan told Reuters the government would consider statutory pre-deployment testing requirements if voluntary frameworks prove insufficient following the AISI containment failures. The statement is the first direct ministerial link between the sandbox breach incidents and potential mandatory legislation.