The European Parliament has reached a consensus on a legislation aimed at governing artificial intelligence (AI). The law includes a prohibition on the utilization of applications linked to significant security risks, such as facial recognition. Esteemed experts in the field have expressed their approval of this momentous decision.
The European Parliament has successfully enacted the world’s inaugural legislation on artificial intelligence (AI). The Members of the European Parliament (MEPs) have endorsed a comprehensive document that will undergo subsequent negotiations with member states and the EU Commission in the following months. Known as the “AI Act,” this legislation prohibits the usage of high-risk AI systems, including real-time biometric facial recognition in public areas. German MEP Svenja Hahn (FDP) firmly stated, “Facial recognition for surveillance, as seen in China, has no rightful position within a liberal democracy.”
Furthermore, there are plans to prohibit or impose strict limitations on other applications that pose significant risks to human safety. The requirements will be classified based on the level of risk associated with each application. The categories include low-risk, moderately risky, highly risky, and prohibited AI. In line with the intentions of the European Parliament, AI systems that classify individuals based on their social behavior or ethnic traits would not be permissible.
Expert Cautions Against Excessive Regulation
Digital association Bitkom expresses its initial approval of the law, according to Achim Berg, the head of the association. Nevertheless, it is crucial to prevent excessive regulation during further deliberations. Berg emphasizes the need for a precise and narrow definition of “high-risk application.” He suggests that regulations should focus solely on applications that genuinely entail significant hazards and risks in practical use.
Additionally, Berg calls for a clear delineation of AI within the “AI Act,” cautioning against transforming it into a broad “Software Act” that regulates all statistical systems or virtually every software. Sam Altman, CEO of OpenAI, the company behind ChatGPT’s development, has also cautioned against the existential dangers of unregulated AI. On the other hand, smaller developers express concerns about their ability to comply with extensive documentation requirements. Civil rights activists hope that the AI law will effectively prevent discrimination and minimize disadvantages.
Advertisement
Potential Implementation by 2026
In 2021, the EU Commission presented its proposal for the first-ever global AI regulation law. With the Parliament now reaching a consensus, negotiations will commence with EU member states and the Commission to finalize the law’s wording. If an agreement can be reached before the upcoming European elections, it is likely that the AI regulation will be implemented by 2026. It is worth noting that artificial intelligence will continue to advance during this period, implying the likelihood of additional AI regulations within the European Union in the future.
An Introduction to Artificial Intelligence
Artificial intelligence primarily pertains to machine learning-based applications, where software efficiently analyzes vast datasets, identifies patterns, and draws meaningful conclusions. These applications are already extensively employed across various domains. For instance, they enable rapid and precise evaluation of computer tomography images, surpassing human capabilities. Self-driving cars employ similar techniques to anticipate the behavior of other drivers on the road. Additionally, AI powers chatbots and curated playlists on streaming services, enhancing user experiences.
