The European Parliament has approved its stance on the AI Act.


The European Parliament has achieved a major milestone in the regulation of artificial intelligence by overwhelmingly voting in favor of adopting its position on the forthcoming AI Act. This act aims to govern AI by taking into account its potential for harm and adopting a risk-based approach. It seeks to prohibit applications that pose an unacceptable level of risk while implementing strict regulations for high-risk use cases.

The timing of AI regulation has been a topic of debate, but Dragoș Tudorache, one of the European Parliament's co-rapporteurs on the AI Act, emphasized the importance of regulating AI at this juncture due to its profound impact.

The adoption of the AI Act encountered uncertainties as a political agreement fell apart, leading to amendments proposed by various political groups. One contentious issue was the use of Remote Biometric Identification, with liberal and progressive lawmakers aiming to restrict its real-time use except for post-investigation of serious crimes. The center-right European People's Party attempted to introduce exceptions for exceptional circumstances such as terrorist attacks or missing persons, but their efforts were unsuccessful.

The AI Act introduces a tiered approach for AI models, with stricter regulations applying to foundation models and generative AI. In line with the act, the European Parliament plans to implement mandatory labeling for AI-generated content and require the disclosure of training data protected by copyright. This step is a response to the growing attention on generative AI, like ChatGPT, prompting the European Commission to launch initiatives for international alignment on AI regulations.

MEPs made several significant amendments to the AI Act, expanding the list of prohibited practices to include subliminal techniques, biometric categorization, predictive policing, internet-scraped facial recognition databases, and emotion recognition software. Furthermore, an additional layer of regulation was introduced for high-risk AI applications, extending the list of high-risk areas and use cases to encompass law enforcement, migration control, and recommender systems on major social media platforms.

Robin Röhm, CEO of Apheris, acknowledged the passing of the plenary vote as a significant milestone in AI regulation but also expressed concerns about the potential burden it may impose on startups and investors, potentially making it less attractive to deploy capital into companies operating in the EU.

With the European Parliament adopting its position on the AI Act, interinstitutional negotiations will commence with the EU Council of Ministers and the European Commission. These negotiations, known as trilogues, will address key points of contention such as high-risk categories, fundamental rights, and foundation models.

Spain, which will assume the rotating presidency of the Council in July, has prioritized the finalization of the AI law as its top digital objective. The aim is to reach a deal by November, with multiple trilogues planned as a backup.

As the negotiations progress, the EU is expected to intensify its efforts to establish comprehensive regulations for AI that strike a balance between innovation, governance, and the protection of fundamental rights. Overall, the European Parliament's adoption of its position on the AI Act represents a significant step towards regulating artificial intelligence, bringing the EU closer to a comprehensive framework.