European Parliament Approves Landmark EU AI Act, Paving Way for Global AI Regulation
Summary:
The European Parliament has approved the EU AI Act, one of the first comprehensive AI regulations worldwide. The Act aims to ensure safety, respect for fundamental rights and innovation support. It categorizes AI systems based on the risk they pose, from 'unacceptable' to 'minimal'. The legislation requires AI developers to provide detailed summaries of their training data for system development and adhere to EU copyright law. The act, initially facing pushback from tech companies, recently received praise from IBM. It symbolizes a step forward in fostering open and trustworthy AI ecosystems.
On March 14, the European Union's AI act became law with the explicit approval of the European Parliament, pioneering comprehensive AI regulation globally. The law, applicable across 27 EU member states, aims to ensure AI technology is both safe and respects fundamental rights within the EU, whilst simultaneously fostering innovation, as stated on the EU Parliament's website. The legislative endorsement secured 523 votes for, 46 against and 49 abstentions.
Before casting their votes, a virtual press briefing was held, where EU Parliament members Brando Benifei and Dragos Tudorache expressed their views to the media, labeling it a momentous day in the journey towards AI regulation. Benifei stressed that the law would allow the development of AI that prioritizes safety and a human-centric approach, as per the priority areas outlined by the EU Parliament.
This legislation, initially proposed five years ago, gained momentum in the past year coinciding with the development and mass deployment of influential AI models. After lengthy negotiations, Parliament arrived at a provisional agreement in December 2023. Subsequently, the agreement was endorsed 71-8 by the Internal Market and Civil Liberties Committees on February 13. As the final voting took place, Tudorache emphasized the importance of international cooperation in AI governance.
Following this legislative action, minor linguistic adjustments will be made during the law's translation phase that ensures understanding in all member states' languages. A second voting procedure for the bill will then occur in April, with its publication in the official EU journal likely to follow in May, as reported by EuroNews. From November onwards, bans on prohibited practices will begin to be enforced, as confirmed by Benifei.
The EU AI Act classifies machine learning models into four categories according to societal risk levels, with high-risk models subject to stricter regulations. Topmost category, 'unacceptable risk', bans AI systems posing threats to safety, livelihoods, and people's rights. This includes governmental social scoring to voice-assistant toys prompting risky behavior.
High-risk applications comprise critical infrastructure, education and vocational training, product safety components and law enforcement among others, that interfere with individuals' fundamental rights. The 'limited risk' category covers transparency in AI usage with examples like interaction with AI chatbots and the mandatory identification of AI-generated content. To assist organizations in compliance, the EU dveloped a tool, 'The EU AI Act Compliance Checker'.
The law also provides for the 'free use' of 'minimal-risk' AI applications, including AI-enabled video games and spam filters, which according to the EU, currently forms the overwhelming majority of AI systems in use in the EU.
Generative-AI models received additional provisions due to their popularity and the accessibility of AI chatbots including ChatGPT, Grok, and Gemini. Developers of general-purpose AI models are required to submit detailed summaries of their training data used for system development and meet EU copyright law.
Although the EU's AI Act faced previous backlash from local businesses and tech companies cautioning over excessive regulation impeding innovation, it received recent praise from industry giant IBM. Christina Montgomery, IBM's Vice President and Chief Privacy and Trust Officer, lauded the EU for passing smart, comprehensive AI legislation that balances risk, aligns with IBM's ethical AI practices, and fosters open and trustworthy AI ecosystems.
Published At
3/13/2024 2:57:48 PM
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.