The European Parliament has taken a significant step towards the regulation of artificial intelligence by voting to adopt its position for the upcoming AI Act with an overwhelming majority.
The act aims to regulate AI based on its potential to cause harm and follows a risk-based approach, prohibiting applications that pose an unacceptable risk while imposing strict regulations for high-risk use cases.
The timing of AI regulation has been a subject of debate, but Dragoș Tudorache, one of the European Parliament’s co-rapporteurs on the AI Act, emphasised that it is the right time to regulate AI due to its profound impact.
Dr Ventsislav Ivanov, AI Expert and Lecturer at Oxford Business College, said: “Regulating artificial intelligence is one of the most important political challenges of our time, and the EU should be congratulated for attempting to tame the risks associated with technologies that are already revolutionising our daily lives.
“As the chaos and controversy accompanying this vote show, this will be not an easy feat. Taking on the global tech companies and other interested parties will be akin to Hercules battling the seven-headed hydra.”
The adoption of the AI Act faced uncertainty as a political deal crumbled, leading to amendments from various political groups.
One of the main points of contention was the use of Remote Biometric Identification, with liberal and progressive lawmakers seeking to ban its real-time use except for ex-post investigations of serious crimes. The centre-right European People’s Party attempted to introduce exceptions for exceptional circumstances like terrorist attacks or missing persons, but their efforts were unsuccessful.
A tiered approach for AI models will be introduced with the act, including stricter regulations for foundation models and generative AI.
The European Parliament intends to introduce mandatory labelling for AI-generated content and mandate the disclosure of training data covered by copyright. This move comes as generative AI, exemplified by ChatGPT, gained widespread attention—prompting the European Commission to launch outreach initiatives to foster international alignment on AI rules.
MEPs made several significant changes to the AI Act, including expanding the list of prohibited practices to include subliminal techniques, biometric categorisation, predictive policing, internet-scraped facial recognition databases, and emotion recognition software.
An extra layer was introduced for high-risk AI applications and extended the list of high-risk areas and use cases in law enforcement, migration control, and recommender systems of prominent social media platforms.
Robin Röhm, CEO of Apheris, commented: “The passing of the plenary vote on the EU’s AI Act marks a significant milestone in AI regulation, but raises more questions than it answers. It will make it more difficult for start-ups to compete and means that investors are less likely to deploy capital into companies operating in the EU.
“It is critical that we allow for capital to flow to businesses, given the cost of building AI technology, but the risk-based approach to regulation proposed by the EU is likely to lead to a lot of extra burden for the European ecosystem and will make investing less attractive.”
With the European Parliament’s adoption of its position on the AI Act, interinstitutional negotiations will commence with the EU Council of Ministers and the European Commission. The negotiations – known as trilogues – will address key points of contention such as high-risk categories, fundamental rights, and foundation models.
Spain, which assumes the rotating presidency of the Council in July, has made finalising the AI law its top digital priority. The aim is to reach a deal by November, with multiple trilogues planned as a backup.
The negotiations are expected to intensify in the coming months as the EU seeks to establish comprehensive regulations for AI, balancing innovation and governance while ensuring the protection of fundamental rights.
“The key to good regulation is ensuring that safety concerns are addressed while not stifling innovation. It remains to be seen whether the EU can achieve this,” concludes Röhm.
(Image Credit: European Union 2023 / Mathieu Cugnot)
Similar: UK will host global AI summit to address potential risks
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The event is co-located with Digital Transformation Week.