EU's AI Act Gets Final Approval

0:00

/1861

1×

Facts

• The EU's Artificial Intelligence (AI) Act will officially take effect next month after the bloc's member countries voted Tuesday in favor of a December deal to approve legislation drafted by the European Commission.¹
• The European Council's approval means the law will begin to come into force in 20 days, with phased implementation taking over two years to complete.²
• The act, among other policies, bans AI for cognitive manipulation, social scoring, predictive policing, and categorization by race, religion, or sexual orientation within the EU.³
• Calling it the 'first of its kind in the world,' the Council claims the law sets a contextual 'risk-based' approach, with 'limited risk' and 'high-risk' systems holding varying transparency obligations.⁴
• The European Council continued that it was committed to 'foster[ing] the development and uptake of safe and trustworthy AI systems.' The AI Act holds exemptions for military, defense, and research purposes.⁴
• Penalties for breaches vary between €7.5M ($8.14M) or 1.5% of turnover, and €35M ($38M) or 7% of global turnover, based upon the nature of the legislation's variation.⁵

Sources: ¹reuters.com, ²TechCrunch, ³XINHUA, ⁴consilium.europa.eu and ⁵The Northern Daily Leader.

Narratives

• Narrative A, as provided by ft.com. Once again, the EU has set world-leading regulatory standards, acting quickly and in unison in a unique manner within international politics. Though the EU’s AI regulation may not become a universal model, it provides a commendable framework for other governments to work toward. Despite critiques and challenges, the EU leads in methodically addressing AI's risks and benefits.
• Narrative B, as provided by www.euractiv.com. While a step in the right direction, the EU's AI Act contains limitations that must be addressed. Despite some successes, the legislation offers exemptions to industries both in the private and public sectors that are most in need of restriction and scrutiny. For now, if the law is to succeed and European democracy is to be protected, close coordination with civil society and human rights experts is a necessity.
• Narrative C, as provided by Reason.com. EU lawmakers have chosen to stifle the innovation of technology they do not understand. Despite good intentions, governmental AI regulation is dependent on a level of knowledge that EU bureaucrats lack, and it is likely that the law will only become a detriment to the bloc's future success.

Predictions