The European Parliament (EP) adopted a negotiating position with a large majority during its plenary session on Wednesday regarding the regulation of artificial intelligence (AI). In this position, the Members of Parliament emphasise that the development and application of artificial intelligence within Europe must comply with European fundamental rights and values.
The rules approved by the EP establish obligations for service providers and users based on the level of risks involved. Prohibition would apply to AI systems that pose an unacceptable level of risk to human safety,
such as social credit systems, as well as to AI applications that infringe upon privacy and engage in discriminatory practices.
These include real-time and retrospective remote biometric identification in public spaces, biometric categorisation of sensitive personal data, predictive law enforcement based on profiling or residence and criminal records, the use of emotion recognition in crime prevention, border control, workplaces, and educational institutions, as well as the creation of facial recognition databases using non-targeted facial images from the internet or closed-circuit television networks. The MEPs have classified AI systems that significantly endanger human health, safety, fundamental rights, or the environment as high-risk. AI systems capable of influencing election results and voters, as well as recommendation systems used by social media platforms with over 45 million users are also included on this list according to the proposal put forth by the members of parliament.
According to the EP’s position, developers of general-purpose AI systems would only be able to market their products on the EU market after assessing and mitigating potential risks (to health, safety, fundamental rights, the natural environment, democracy, and the rule of law) and registering their models in the EU database. Transparent labelling would be expected for such models, including systems based on generative AI like ChatGPT, indicating that the content was produced by artificial intelligence. Assistance would be provided to distinguish so-called ‘deepfake’ images from real ones, and appropriate safeguards would be put in place to prevent the generation of illegal content. To support innovation and small and medium-sized enterprises, the MEPs suggested that not all obligations should apply to research activities and AI components made available under open-source licenses. Additionally, the EP advocates for citizens to have the opportunity to lodge complaints regarding the use of AI systems, and should receive detailed information about decisions that significantly impact their fundamental rights if those decisions involve high-risk AI systems.
According to Member of the European Parliament Edina Tóth, ‘artificial intelligence has numerous application areas and holds many possibilities, but it is obvious that AI cannot remain unregulated.’ The representative highlighted the importance of transparency and traceability. ‘We support innovation, but we also need to consider the potential dangers of its application and regulate them with due care. I trust that this legislation can provide the legal framework for artificial intelligence to function as a safe tool for citizens and businesses. The fact that my speech during the parliamentary debate was also written by artificial intelligence demonstrates the advancement of AI systems,’ the politician emphasised.
Fidesz Member of the European Parliament Balázs Hidvéghi emphasised that the vote is an important milestone in shaping regulations that can address citizens’ fears and risks associated with artificial intelligence while avoiding overregulation and hindering innovation. He added that the aim should be to achieve a final regulation that is
human-centred and prioritises consumer protection.
‘To increase citizen trust in AI, we must first identify and filter out the risks. We should favour usage methods that benefit individuals and society. It is encouraging that European regulations also follow this approach,’ the representative said.