EU adopts a new AI law
By St Fox / March 22, 2024
Unveiling the future
Our experts analyze the latest tech trends and industry breakthroughs.
EU adopts a new AI law
Highlights of the new law adopted by the European parliament around AI
The Artificial Intelligence Act, championed by Members of the European Parliament (MEPs), establishes comprehensive rules aiming to ensure AI technologies are developed and used in safe ways, respect fundamental rights, democracy, the rule of law, and environmental sustainability, while also fostering innovation and positioning Europe as a leader in this domain. Here are some of the highlights of the Act based on the provisional agreement reached between Parliament and Council negotiators:
1. Safeguards for General Purpose AI:
The Act mandates that general-purpose AI (GPAI) systems, and the models they are based on, adhere to transparency requirements including technical documentation, compliance with EU copyright law, and the dissemination of detailed summaries about the content used for training. High-impact GPAI models with systemic risk are subject to stricter obligations, such as model evaluations, systemic risk assessments, adversarial testing, serious incident reporting, cybersecurity assurance, and energy efficiency reporting.
2. Limitations and Bans:
There are significant restrictions on certain uses of AI, including the limitation of biometric identification systems by law enforcement, and outright bans on social scoring, AI that manipulates or exploits vulnerabilities, emotion recognition in workplaces and educational settings, and the untargeted scraping of facial images to create recognition databases.
3. Law Enforcement Exemptions:
While establishing stringent safeguards, the Act allows for narrow exceptions for law enforcement uses of biometric identification in publicly accessible spaces, subject to prior judicial authorization and strictly defined criteria related to serious crimes.
4. Obligations for High-Risk Systems:
AI systems classified as high-risk must fulfill clear obligations, including mandatory fundamental rights impact assessments. These systems include those used in influencing elections, voter behavior, and those with significant potential harm to health, safety, fundamental rights, and the environment.
5. Support for Innovation and SMEs:
The Act encourages the development of AI through regulatory sandboxes and real-world testing, particularly benefiting SMEs by promoting innovation without undue pressure from industry giants.
6. Sanctions for Non-Compliance:
Companies not adhering to the Act's stipulations face fines ranging from 35 million euros or 7% of global turnover to 7.5 million euros or 1.5% of turnover, based on the infringement's nature and the company's size.