Sapiens
Policy

What is AI regulation?

Published June 1, 2026 · 5 min read

AI REGULATION · RISK TIERSMore potential harm, more rules.The dangerous tip is banned; rules ease toward the broad, harmless base.Unacceptablebanned outrightHigh-riskheavily regulatedLimited-riskjust disclose it's AIMinimal-riskfree to usemore harm = more rules

Definition

The laws governing how organizations build, sell, and use AI, with stricter duties for riskier uses.

At a glance

  • Risk-based: the EU AI Act sorts AI into four tiers, banned to unregulated[1].
  • Reaches across borders: fines up to 35M euros or 7% of global turnover[3].
  • You have duties even if you only use AI, not build it.
  • The US has no federal law, just a patchwork of state rules[4].

How the tiers work

The EU ranks AI by potential harm. Unacceptable uses (social scoring, manipulation) are banned. High-risk (hiring, lending, medical) is allowed but tightly regulated: human oversight, documentation, registration[2]. Limited-risk just needs disclosure (“you’re talking to a bot”). The rest is minimal-risk and free.

What businesses must do

Map where AI touches real decisions about people. Deploy a high-risk vendor system, and you must keep a human in the loop and disclose its use[2]. EU deadlines stagger: bans hit Feb 2025, most high-risk duties Aug 2026[3].

US picture

States moved first (Colorado), but a December 2025 federal order now seeks to override conflicting state rules, so watch both levels[5].

Bottom line

The more a tool can hurt someone, the more rules apply, up to a ban, with the EU leading across borders and the US a moving patchwork.

References

  1. High-level summary of the AI Act. Future of Life Institute (EU Artificial Intelligence Act) artificialintelligenceact.eu
  2. AI Act | Shaping Europe's digital future. European Commission digital-strategy.ec.europa.eu
  3. U.S. Companies Face EU AI Act's Possible August 2026 Compliance Deadline. Holland & Knight www.hklaw.com
  4. State AI Laws - Where Are They Now? Cooley LLP www.cooley.com
  5. New State AI Laws are Effective on January 1, 2026, But a New Executive Order Signals Disruption. King & Spalding www.kslaw.com

Comments

Questions, corrections, and links welcome. Be specific and civil.

  • Loading comments…