EU publishes revised draft of AI regulation

The European Commission published the final text of the draft aimed at regulating artificial intelligence.

The new version does not restrict the powers and independence of the state bodies responsible for protecting fundamental rights and freedoms. This means the draft does not prohibit the use of AI for national security purposes.

The authors of the document classified risk assessment and pricing algorithms for health and life insurance as high-risk. The rest remain under sectoral legislation.

Algorithms for evaluating credit scores or insurance contributions were also deemed high-risk. However micro and small enterprises deploying these systems for ‘own use’ were exempted from stringent obligations.

Lawmakers believe that the amendment will close a loophole used by large companies via small operators.

The text of the updated regulation has already been sent to EU member states. Its discussion is scheduled for November 18.

“The Committee of Permanent Representatives is invited to consider and approve the text set out in the Annex to this draft with a view to developing a common approach,” the document states.

In addition, the European Commission pledged to issue guidance on applying the provisions of the AI regulation.

Earlier, political groups in the European Parliament submitted several thousand amendments to the draft. Disagreements concerned the definition of artificial intelligence, the use of biometric identification systems and the classification of high-risk algorithms.

In April 2021, the European Commission introduced a proposal to regulate AI.

In June, a group of rights advocates criticized the document and called for banning the use of tracking algorithms, including facial recognition.

Subscribe to ForkLog news on Telegram: ForkLog AI — all the news from the world of AI!