EU Reaches Preliminary Agreement on Comprehensive Artificial Intelligence Regulations
Summary:
The EU Parliament and Council have reached a provisional agreement on the governance of artificial intelligence (AI), covering areas like governmental use of AI in biometric surveillance, regulation of AI systems, and transparency requirements. The EU aims to be the first international authority with laws on AI usage, aiming to highlight its benefits while mitigating risks. The agreement also restricts certain uses of AI like behavioral manipulation, facial data scraping, and social scoring. Now, it awaits formal adoption by the Parliament and Council to become EU law.
On Friday, December 8, the European Union Parliament and Council representatives came to a preliminary agreement on regulations governing artificial intelligence (AI). The consensus discusses how governmental entities should use AI in biometric surveillance, the rules for regulating AI systems like ChatGPT, and the transparency procedures that need to be adhered to prior to market introduction. This includes technical paperwork, compliance with EU copyright laws, and the sharing of training content summaries. The objective of the EU is to lead the way as the first international authority to create AI law, outlining its beneficial application while instituting protection against potential hazards. Following nearly a 24-hour debate on December 8, and an additional 15 hours of negotiations, the agreement was finalized.
The consensus states that AI models with major impact and systemic risk need to evaluate and manage those risk factors, undergo adversarial tests for system stability, report any incidents to the European Commission, assure cybersecurity, and disclose their energy efficiency. The European Parliament will vigilantly monitor these developments, with the aim of fostering new business concepts through sandboxes and the implementation of efficient rules for the most powerful AI models. Following the agreement, Thierry Breton, European Commissioner, expressed his fulfillment on X (formerly called Twitter), proclaiming it a ground-breaking occasion and viewing the AIAct as a springboard for EU startups and researchers to pave the way in the global AI contest.
The agreement also requires general-purpose artificial intelligence systems with risk factors to follow prescribed codes. The use of real-time biometric surveillance by governments is restricted, only permitting its use in distinct incidents such as grave threats in public areas or specific criminal cases. Additionally, the deal prohibits cognitive behavioral manipulation, the scraping of facial images from the internet or CCTV footage, social scoring, and biometric systems deducing personal information like beliefs and orientation. Consumers will have the right to lodge grievances and seek clarifications. Fine penalties for infractions range from 7.5 million euros or 1.5% of turnover to 35 million euros or 7% of global turnover. These penalties will depend on the violation severity and the size of the company committing the infringement.
The final adoption of the agreed text by the Parliament and Council is still pending before it can be translated into EU law. The Parliament’s Internal Market and Civil Liberties committees are expected to cast their vote on the agreement at an impending meeting.
Published At
12/9/2023 11:46:29 AM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.