AI governance & legal landscape
Key Questions
What new AI safety law was passed in Illinois?
Illinois SB 315 mandates third-party audits for certain AI systems to address safety and accountability.
When is the EU AI Act compliance deadline?
The EU AI Act compliance deadline is 30 days away, requiring organizations to meet new transparency and risk obligations.
How is New York using AI for legislation?
NY's governor deployed AI to scan 18 million words of state statutes for review and potential updates.
What accountability framework applies to AI agents?
Frameworks reference Moffatt v. Air Canada and the EU Product Liability Directive, placing responsibility on deploying companies.
What governance gaps exist for frontier models?
Government processes for approving frontier model releases remain opaque, raising transparency concerns.
How should enterprises govern GenAI before scaling?
Recommendations include implementing routing, pseudonymization, and LLM gateways to close data governance gaps.
What initiative did the ITU launch for AI agents?
The ITU launched a trust initiative to improve reliability and accountability of autonomous AI agents.
What legal action involves Apple and OpenAI?
Apple sued OpenAI alleging trade secret theft related to AI technology.
Developing: Illinois AI Safety Law (SB 315) with mandatory third-party audits. EU AI Act compliance deadline in 30 days. NY governor using AI to scan 18M words of state statute. Stanford HAI research on AI hiring bias. Government opacity on frontier model approval process. ITU trust initiative for AI agents. Apple sues OpenAI for trade secret theft. Swarm Accountability Gap paper. OpenAI safety lead departure. Orca report highlights systemic security hygiene crisis as governance gap. New: 'creeping normality' of LLMs in life sciences — subtle shift before proper guardrails agreed upon. AI Ethics 2.0 for healthcare (dynamic institutional oversight), AI agent accountability framework (Moffatt v. Air Canada precedent, EU Product Liability Directive), LLMs amplify existing data governance gaps faster than previous enterprise software; enterprises need routing, pseudonymization, and LLM gateways before scaling gen AI.