AI Act: an agreement was reached

11/12/23

On the evening of 8 December, the European Council presidency and the European Parliament’s negotiators, alongside members of the European Commission, have reached a provisional agreement on the AI Act, a bill that prioritises safety, funda...

Share this news
Stay up to date!

Follow us on Twitter

Follow us on Linkedin

Subscribe to our monthly Newsletter!

On the evening of 8 December, the European Council presidency and the European Parliament’s negotiators, alongside members of the European Commission, have reached a provisional agreement on the AI Act, a bill that prioritises safety, fundamental rights and democracy while stimulating innovation and nurturing business growth on a European level. The landmark ‘Artificial Intelligence Act’ is the first well-developed act that deals with AI, carefully analysing implications related to privacy and individual rights, implications for businesses and society and represents a cornerstone for world AI legislation.

The deal, reached after 3 days of intensive negociations, introduces regulations focusing on high-risk AI, governance and prohibited practices. It defines AI systems, identifies high-risk categories and imposes obligations accordingly. Notably, it bans certain harmful AI practices like cognitive manipulation and emotion recognition in sensitive areas like workplaces and education.

Law enforcement exceptions are outlined with safeguards to preserve confidentiality and protect fundamental rights. Additionally, the agreement addresses general-purpose AI systems and introduces a governance structure involving an AI Office, a scientific panel and an advisory forum.

Penalties for violations are set as a percentage of a company’s turnover, with capped fines for SMEs and startups. Fundamental rights assessments and increased transparency requirements for high-risk AI systems are included, along with measures to support innovation through regulatory sandboxes and real-world testing.

BDVA’s role

In this historic moment, BDVA is honoured to offer its support on the implementation of the AI Act, working alongside European industries and research organisations. One of BDVA’s task forces, etami, champions and advances trustworthy and ethical design, development and deployment of AI systems. It accomplishes this by translating European and global ethical AI principles into practical and quantifiable tools and methods. These encompass a spectrum of resources, ranging from checklists to algorithmic auditing. etami’s focus spans across various work streams, delving into the foundational aspects of ethical and dependable AI, including AI and law, analysis of AI incidents database, AI life cycle models, AI auditing/ML testing, transparent AI or AI explainability, AI security, AI privacy, AI fairness, documentation practices for AI systems and data quality.

What will happen next?

The agreement is subject to technical refinements before submission to member states for endorsement. Once finalised and legally-linguistically revised, it will undergo formal adoption by European institutions.

Background

This proposal, originating in April 2021, aligns with the EU’s strategy to foster safe AI adoption, enhance governance and establish a unified market for AI applications. It follows a risk-based approach and complements initiatives to accelerate AI investment in Europe. The Council agreed on a negotiating mandate in December 2022, initiating interinstitutional talks with the European Parliament in June 2023.

For more information: https://www.europarl.europa.eu/news/en/press-room/20231206IPR15699/artificial-intelligence-act-deal-on-comprehensive-rules-for-trustworthy-ai

https://www.consilium.europa.eu/en/press/press-releases/2023/12/09/artificial-intelligence-act-council-and-parliament-strike-a-deal-on-the-first-worldwide-rules-for-ai/

Make sure to visit and follow etami.org!