Published in News

The first part of EU’s new AI law running

by on03 February 2025


Apple and Meta already unhappy about where this is going

The first requirements of the European Union’s AI Act have now come into force, effectively banning artificial intelligence systems deemed "unacceptably risky."

EU regulators now have the authority to eliminate any dodgy AI products and companies that attempt to violate these rules face potential fines of up to €35 million (around $36 million) or seven per cent of their global revenue—whichever amount is greater.

The Act was approved by the European Parliament in March last year after extensive deliberation and targets AI systems that pose significant risks to individual rights and societal values. It explicitly prohibits systems like China’s social credit system, which adjusts individuals' credit ratings based on behaviour or reputation.

Also banned are AI systems using subliminal messaging to manipulate choices and profile vulnerable populations such as minors or persons with disabilities. Law enforcement tools predicting criminal behaviour based on facial features are all banned.

Real-time monitoring systems for law enforcement are more tightly regulated and permissible only in specific scenarios. This means police cannot deploy facial recognition at public events or in subway stations to identify suspects.

The Act also outlaws AI systems that aggregate biometric data to infer political beliefs, gender, or sexual orientation, as well as "emotion-tracking" technologies—except in cases related to medical treatment or safety.

These bans apply to any company operating within EU member states, including non-EU entities offering services to EU citizens. The Act represents the most substantial governmental effort to regulate AI to date. Many US tech giants have expressed willingness to comply—over 100 organizations, including Google, Amazon, Microsoft, and OpenAI, signed the voluntary "EU AI Pact."

However, some notable companies have abstained from the voluntary pact. Meta, Apple and French startup Mistral argue that the regulations are overly rigid and may hinder innovation within the AI industry.

Despite their refusal to sign the pact, these companies are not exempt from the new law and will face significant fines if they operate AI systems that contravene EU rules.

Rate this item
(0 votes)