The world's first comprehensive supervision AI EU version of the ＂Artificial Intelligence Act＂
The一刻资讯 European Parliament, EU member states, and the European Union Commission finally reached an agreement on the "Artificial Intelligence Act" on the evening of December 8, local time in Brussels.
This bill will become the world's first comprehensive regulatory regulations in the field of artificial intelligence.
According to Reuters, the bill will take effect early next year.
EU's internal market commissioner Tiery Brendon admits that the agreement will help Europe "lead the global artificial intelligence competition."
"Artificial Intelligence Act"
The bill is classified according to different risk categories for the application of artificial intelligence technology. From "unacceptable", that is, technologies that must be prohibited, to high, medium, and low -risk artificial intelligence, etc., are supervised by identifying different risks.
Strictly prohibit the "artificial intelligence system that causes unacceptable risks to human security", including systems that are managed by destinations, use human weaknesses, or evaluate systems such as behavior, social status and personal characteristics.
Artificial intelligence companies should control their algorithms, provide technical documents, and establish a risk management system for "high risk" applications.Each EU member states will set up a supervision agency to ensure that these rules are observed.
Generating artificial intelligence, which is a model including ChatGPT, also needs to comply with new transparency requirements.The European Union requires it to draft technical documents before launching the market and abide by the EU copyright law.
A tug -of -war
Earlier, the European Union has always hoped to promote the "Artificial Intelligence Act" as soon as possible.
Since 2019, the European Union has been deploying the field of artificial intelligence supervision, and has released artificial intelligence ethics standards and introduced the "Artificial Intelligence White Paper".
In April 2021, the European Commission proposed a draft of negotiating authorization of the "Artificial Intelligence Act" proposal.Since then, the European Parliament and the EU Council have revised and discussed several rounds of revisions on drafts.
In June 2023, the European Parliament used 499 votes, 28 votes against and 93 abstrams. High votes passed the draft of the "Artificial Intelligence Act" negotiations and promoted the final stage of the bill to enter the legislative procedure, that is, the European Parliament and the European Union were held.The "three parties negotiations" of member states and the European Commission to determine the final terms of the bill.
Last week, the three parties started a marathon -style closed -door negotiation. One of the negotiations lasted 22 hours, and the second round of negotiations started on the morning of December 8.
Camegas, a Secretary of State, Secretary of the Secretary to the negotiations, said that the effort to ensure the "milestone" supervision is worth it.
Some EU officials have expressed their hope that the new bill can help the EU region's artificial intelligence technology better develop and enhance the competitiveness of the industry.
Blaidon, the internal market commissioner of the European Commission, said: I believe that the "Artificial Intelligence Act" is not only a regular manual, but also a booster for the European Union startup and researchers to lead the global artificial intelligence competition.
"The result of the compromise of all parties"
Wu Shenfang, a doctoral supervisor instructor in the School of Law of Beijing Normal University and deputy director of the Research Center of the China Internet Association, said that the EU parties finally reached an agreement on the Artificial Intelligence Act, which is the result of compromise in all parties.Demand also has pressure from the external environment.
First, the EU countries agree that the development of artificial intelligence can enable the EU's competitive advantage and win the right to speak in the future global governance, and countries also have practical needs of industrial and economic development.Therefore, the European Union is urgent for artificial intelligence legislation.
Second, international governance subjects including China, the United States, and the United Kingdom are rapidly promoting the rules of artificial intelligence governance.In terms of external environment, it also determines that the EU needs to form and launch its own rules as soon as possible.
Wu Shenjuo further analyzed that the EU's "Artificial Intelligence Act" will have a very significant impact on the artificial intelligence development pattern and legislative pattern of the European Union and the world.
The legislative skills of the Artificial Intelligence Act are comprehensive and comprehensive.The promulgation and implementation of the future bill will have a great direct impact on the design and operation of artificial intelligence, the design and operation of artificial intelligence products and services.
At the level of compliance, the bill will affect the market structure, business model, operation method and even development prospects of various artificial intelligence companies.From the perspective of law enforcement and implementation, it is a relatively large challenge for artificial intelligence companies, and it is also a process of business reconstruction and business design.
After the bill is officially promulgated and implemented, the member states will make certain changes based on the foundation of the country's public policy.
What is intriguing is that before, dozens of large European companies including Airbus and Siemens have adopted collective operations on the grounds of "may damage European competitiveness" and "cannot cope with potential challenges", and publicly opposed the EU through "Artificial Intelligencebill".