European Union agrees on draft of world’s first AI regulation law

The European Union (EU) has agreed on a draft bill to regulate artificial intelligence (AI) for the first time in the world.

According to foreign media such as the New York Times on the 8th (local time), the EU Parliament and member states agreed to introduce an AI regulation law in Brussels, Belgium. European Commissioner Thierry Britton even left a short note on his “Europe has established itself as a pioneer by understanding the importance of its role as a global standard-setter,” Britton said in a statement.

The bill, called the AI ​​Regulatory Act, seeks to leverage the technology’s potential benefits while protecting against potential risks such as automation, the spread of misinformation, and national security threats. The New York Times explained, “This bill still requires final approval, but this agreement means that the outline of the bill has been set.”

Europe has focused on the risky use of AI by companies and governments. Not only must creators of large-scale general-purpose AI systems such as ChatGPT meet requirements related to transparency, but creators of deepfake software (SW), including chatbots, must clearly disclose that the output of their software was created by AI. . Facial recognition software is restricted except in certain circumstances such as security, and companies that violate these regulations can be fined up to 7% of their sales.

However, the New York Times said, “This law is expected to take effect in 12 to 24 months, which is a fairly long time to develop AI,” and added, “There are also concerns raised about the limits to the effectiveness of this law.”

The EU’s AI regulation law is approved through a vote by the European Council, which consists of representatives of 27 member states as well as the Parliament.