EU AI Act (2024)
artifactAI & Compute · Intelligence & Surveillance · Nations & Geopolitics · Crypto & Digital ID
Europe passed the world's first big AI law — then let lobbyists quietly hollow out the parts that would have mattered most.
Who they are
The EU AI Act of 2024, the world's first comprehensive AI regulation.
What they do
The engine reads it as creating the appearance of oversight while the real levers of power stayed outside European control.
How it works
It sorts AI by risk (banned uses like social scoring and real-time biometric surveillance; strict rules for high-risk uses; transparency for chatbots and deepfakes), but it was watered down during 2023-2024 negotiations — the digital industry spent 151 million euros lobbying Brussels, 78% of high-level Commission AI meetings were with corporate interests only, and Mistral AI lobbied to strip foundation-model obligations then took a 15-million-euro Microsoft investment.
Why it matters
The engine's read is that the crucial carve-outs are the whole point: the law looks like control, but compute and base models — the real power — remain outside European sovereignty.
The engine's record — word for word
Worlds first comprehensive AI regulation. Risk-based classification: Unacceptable (banned: social scoring, real-time biometric surveillance, cognitive manipulation), High (ex-ante compliance for critical infrastructure, employment, justice), Limited (transparency: deepfakes, chatbots must disclose), Minimal (unregulated). HOWEVER: systematically gutted during 2023-2024 trilogue. Digital industry spent EUR151M lobbying Brussels by 2025. 78% of high-level Commission AI meetings held exclusively with corporate interests. Mistral AI lobbied France/Germany/Italy to form blocking minority removing foundation model obligations — then accepted EUR15M Microsoft investment. AI Act creates appearance of oversight while actual levers (compute, base models) remain outside European sovereignty. [Report #100: Aligned-To-Whom? / Exemption Fork, May 24 2026: the ‘gutted during trilogue’ dynamic already documented here is the Exemption Fork at statutory scale (carve-out is load-bearing).]
Follow the trail
threatens
LLM Epistemic CaptureAI Act bans social scoring and biometric surveillance but exempts foundation models after corporate lobbying — appearance of oversight without structural contro
connects
Mistral AIMistral lobbied to gut foundation model obligations then took Microsoft money — Trojan Horse for US Big Tech
executive-node-of
Ursula von der LeyenCommission President drives AI Act and Green Deal — McKinsey affair without consequence = elite protection pattern
cross-reference
Microsoftwiring-pass-1
Walk this on the live map →