EU AI Act, US Executive Order 14110, UK pro-innovation framework, China's AI governance rules — four different philosophies, four different compliance costs. Global AI companies face all four simultaneously.
€35MMax EU AI Act fine (or 7% global revenue)
$2.7BEstimated annual global compliance cost for large AI firms
3Years before EU AI Act fully applies (2027)
12+Countries with active AI regulation proposals (2024)
Choose your depth. The data doesn't change — just the explanation.
Different countries are making different rules about AI. Europe has very strict rules and will fine companies millions of dollars if they break them. The UK wants to be more friendly to AI companies. The US is somewhere in the middle. China has its own rules focused on keeping political control. If you're a big AI company, you have to follow ALL of these rules at once, which costs a lot of money.
The EU AI Act (2024) classifies AI systems by risk level — from banned uses (real-time biometric surveillance) to minimal risk — and requires conformity assessments, transparency logs, and human oversight for high-risk uses. US EO 14110 focuses on safety testing and mandatory reporting for frontier models. UK takes a sector-regulator approach. China requires algorithm registration and bans AI that "undermines national unity." Companies operating globally face inconsistent, overlapping, and sometimes contradictory requirements.
EU AI Act (Regulation 2024/1689): risk-based tiering; GPAI models >10²⁵ FLOPs face systemic risk provisions; max fine 7% global annual turnover. US EO 14110: mandatory safety reporting for models trained with >10²⁶ FLOPs, NIST AI Risk Management Framework. UK: "pro-innovation" — no dedicated AI law, existing sector regulators adapt. China: "Interim Measures for Generative AI Services" (2023): requires watermarking, training data disclosure, algorithm registration with CAC. Compliance cost estimate: KPMG (2024) $50-200M/year for large models under EU Act alone; global multi-regime compliance 40-60% premium.
Full texts: EU AI Act (eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689), US EO 14110 (whitehouse.gov), UK DSIT framework (dsit.gov.uk/ai-regulation). NIST AI RMF: nist.gov/system/files/documents/2023/01/26/AI RMF 1.0.pdf. China CAC: cac.gov.cn/2023-07/13/c_1690898327029107.htm. Compliance cost: KPMG "AI Regulation: The Business Impact" 2024. Track: EU AI Act implementation timeline, US NIST AI 100-1 standard, ISO/IEC 42001 (AI Management Systems standard).
The Four Regulatory Regimes
Each jurisdiction takes a fundamentally different approach to AI governance — creating a compliance maze for global companies.
Compliance Cost by Framework and Company Size ($M/year)
KPMG AI Regulation Business Impact Study 2024. Estimates for companies deploying AI in regulated sectors.
"Pro-innovation" — no dedicated AI law. Existing regulators (FCA, ICO, CMA) apply sector-specific rules. Voluntary principles: safety, transparency, fairness.
Existing law applies
China Generative AI Rules
🇨🇳 China — Effective Aug 2023
Algorithm registration with CAC. Watermarking required. Training data disclosure. Prohibits content "undermining national unity." Promotes "core socialist values."
EU regulations historically spread globally because multinationals comply once with the strictest standard. The EU AI Act may become the de facto global standard — with European risk assessment requirements shaping how AI is built everywhere, not just in Europe.
Sources
• EU AI Act (Regulation 2024/1689). eur-lex.europa.eu. Published August 1, 2024.
• Executive Order 14110 on AI Safety. whitehouse.gov. October 30, 2023.
• UK DSIT. "AI Regulation: A Pro-Innovation Approach." dsit.gov.uk. March 2023.
• China CAC. "Interim Measures for Generative AI Services." cac.gov.cn. July 2023.
• KPMG. "AI Regulation: The Business Impact." 2024.