The European Union (EU) has reached a final agreement on the world’s first legislation to regulate artificial intelligence (AI). This regulation primarily targets the use of AI technology in high-risk areas such as law enforcement and employment and is set for full implementation starting in 2026.
On May 21, economic news channel CNBC and German broadcaster Deutsche Welle (DW) reported that the EU Commission had finally approved the AI Act, a comprehensive bill to regulate AI-related technologies.
It was reported that the consensus among EU member countries will become the world’s first major AI regulatory law as the world attempts to regulate the introduction of AI technology. The law requires the AI used in public safety and employment to be sufficiently transparent and accurate and meet cybersecurity regulations.
The draft of the EU’s AI Act was proposed in early 2021 and passed by the European Parliament last March.
Belgian Digital Minister Mathieu Michel said, “With the AI Act, Europe will emphasize trust, transparency, and accountability regarding AI technology. We hope this law will promote both the development of this technology and innovation in Europe.”
According to the regulation, AI apps that can cause risks are banned. This includes social score evaluations through surveillance at workplaces or schools and classifying based on religion, personal beliefs, skin color, and sexual orientation.
Real-time facial recognition through closed circuits can only be used for law enforcement purposes such as finding missing persons, identifying kidnapping victims, preventing human trafficking, and tracking criminal suspects.
The EU has been pushing for a quick revision of the AI law since OpenAI launched the generative AI ChatGPT in November 2022.
The EU AI law, as the world’s first comprehensive AI regulatory law, is expected to significantly influence the construction of AI regulatory models in other countries.
American Big Tech companies have been monitoring the EU’s progress in anticipation of the ripple effects the AI law may bring.
This is because they suspect that the AI law targets American companies and could restrict AI technology development, creation, use, and resale within EU member countries.
The AI law has strong sanctions, allowing the EU Commission to impose a fine on companies of either $38.5 million or 7% of global sales, whichever is higher.
The EU will establish a new AI office to oversee the implementation of the AI Act in compliance with EU requirements.
The AI law will officially take effect next month within the EU’s 27 member countries, and the prohibition on certain AI regulations will be implemented first after six months. A year after that, regulations on general-purpose AI (AGI) will be implemented, with full implementation expected in mid-2026.
Current generative AIs such as ChatGPT, Gemini, and Copilot will be given a grace period of 36 months.
