New AI regulations: what's changing in the EU and US

AI regulation is moving fast. Here’s a concise snapshot of what’s changing.

EU AI Act Risk-based tiers are in force. High-risk applications need documentation and human oversight. General-purpose models face transparency rules. Fines are large; compliance timelines vary by tier.

US State and sector rules are ahead of federal law. Focus areas: bias, transparency, and use in hiring and finance. Expect more federal guidance in 2025.

What to do Map your use case to “high risk” or “general purpose” under the EU Act. Document data, model choice, and human review. Keep an eye on state-level rules if you operate in the US.

Staying compliant is easier if you use established vendors (e.g. OpenAI, Anthropic) that invest in legal alignment.