EU adopted ‘world-first’ AI law
On 13 March 2024 the European Parliament approved the world’s first major set of regulatory ground rules to govern artificial intelligence.
What does the AI Act encompass?
The AI Act introduces a robust framework to regulate AI systems, with a strong emphasis on managing high-risk AI. Key components include:
- Obligations for High-Risk AI
- Prohibited systems
- General-purpose AI (GPAI)
- New organisational framework and penalties
- Implementation timelines
Obligations for High-Risk AI
Organizations deploying AI systems in critical sectors such as healthcare, transportation, and finance will be subject to rigorous obligations. These encompass comprehensive risk assessments, transparency measures, and robust accountability standards.
Prohibited systems
Certain AI applications, such as those involved in manipulating human behaviour or conducting social scoring, will be outright prohibited.
General-purpose AI (GPAI)
GPAI systems, and the GPAI models they are based on, must meet certain transparency requirements, including compliance with EU copyright law and publishing detailed summaries of the content used for training. The more powerful GPAI models that could pose systemic risks will face additional requirements, including performing model evaluations, assessing and mitigating systemic risks, and reporting on incidents. Additionally, artificial or manipulated images, audio or video content („deepfakes“) need to be clearly labelled as such.
New organisational framework
- European AI Office
- European AI Board
- Scientific Panel
- Dedicated National Authorities
- EU Database for High-Risk AI Systems
Penalties
- Up to EUR 35 million or 7% of total worldwide annual turnover
- The AI Act established the European AI Board to oversee compliance
- Penalties for non-compliance will be enforced 12 months after the AI Act enactment
Implementation timelines
The provisions regarding GPAI and penalties will come into effect 12 months after the enactment of the AI Act, while high-risk AI systems must comply within 24 months (36 months for AI systems covered by existing EU products legislation). Prohibited systems will be phased our within 6 months.
The first rules come into force in just a few months.
