With the rapid development of artificial intelligence in recent years, the competition for the dominance of artificial intelligence rule-making in major Western countries has become increasingly fierce. Recently, the European Parliament, EU member states and the European Commission reached an agreement on the "Artificial Intelligence Act", which will become the world's first comprehensive regulation in the field of artificial intelligence. EU Internal Market Commissioner Thierry Breton said that the AI Act is not only a rulebook, but also a booster for EU start-ups and researchers to lead the global AI race.
EU: AI governance legislation first
The EU's approach to AI rule-making is based on legislation, and its legal framework has been gradually improved over the past few years. As early as April 2018, the European Commission issued a document outlining ways to coordinate the development and deployment of AI, with the aim of increasing investment in AI research and innovation. In February 2020, the European Commission issued "Artificial Intelligence*** proposes a regulatory framework for AI, covering issues such as risk assessment, transparency, data use and legal liability, and proposes to classify and regulate AI applications with different risk levels. In April 2021, the European Commission proposed the world's first regulation on artificial intelligence to ensure that AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly. In June 2022, the European Parliament adopted a draft mandate to negotiate proposals for the Artificial Intelligence Act, which classifies the risks of AI systems, restricts deepfakes, and imposes greater transparency requirements for generative AI.
In addition to leading the world in legislative speed, the EU is also trying to implement regulatory standards for AI globally. Brando Benifi, a member of the European Parliament who oversees the work on the AI Act, said EU lawmakers would set a "path for responsible AI" for the rest of the world. According to previous reports, the EU is carrying out a "lobbying blitzkrieg", hoping to persuade Asian countries to recognize the EU's leading position in the field of artificial intelligence regulation, so that the EU "Artificial Intelligence Act" will become a global standard in the field of artificial intelligence regulation. At present, the EU and its member states have sent ** to negotiate with at least 10 Asian countries, including India, Japan, South Korea, Singapore and the Philippines, on this issue.
United States: AI policies encourage the development and application of technology
The U.S. AI governance policy focuses on the development and application of technology, and its AI policy is relatively lenient compared with the EU. Although it also faces the security challenges of artificial intelligence, the United States does not conduct special research and classification of artificial intelligence risks like the European Union, but focuses on the two risk points of fairness caused by algorithmic discrimination and data privacy security protection, in order to maintain the innovation and commercial competitiveness of its own artificial intelligence technology while maintaining fairness and privacy. In terms of governance, the United States tends to address related security risks through industry self-regulation and industry guidelines, such as urging companies to develop their own AI ethics guidelines and reducing the risk of algorithmic discrimination through internal audits and self-monitoring.
In recent years, technologies such as generative AI have matured. In order to ensure the safety and reliability of generative AI technology, the United States** has gradually strengthened supervision and successively issued a series of policy documents to strengthen AI governance, so as to maintain the hegemony of the United States in this field and reduce the security risks of AI. Since the release of the Blueprint for the Artificial Intelligence Bill of Rights in 2022, the United States** has issued a number of principled regulations to guide the design, development, deployment and use of AI systems, and encourage the industry to voluntarily comply with relevant principles and self-regulate, thus gradually forming the basic framework of AI governance in the United States.
Europe and the United States: Rulemaking is both competitive and cooperative
As the "frontrunner" in the field of artificial intelligence rule-making, Europe and the United States have a relationship of competition and cooperation. The cooperation between the two sides is mainly based on the US-EU ** and technical committees established in June 2021 as a platform, and is based on the relevant regulations and policy documents issued by the US and the EU. In December 2022, the US-EU** and the Technical Commission released the Joint Roadmap on Trustworthy AI and Risk Management Assessment and Measurement Tools for the first time, guiding the development of AI risk management and trustworthy AI from the three aspects of terminology specification, standard development and risk monitoring, and promoting the formulation of relevant international standards. In this document, "the United States and the EU jointly support and lead the development of international technical standards" is highlighted.
However, in terms of practical results, although the United States has begun to converge with the EU in the field of AI regulation, there are still some structural problems that are difficult for the two sides to coordinate. For example, at the level of strategic concepts, the U.S. security and foreign policy establishment regards AI as an important asset in great power competition, and it should be used as a tool to expand the influence of technology; However, the EU is more concerned about the ethical challenges of AI technology based on economic development and values. In terms of risk management concepts, the United States encourages the innovation and development of artificial intelligence technology, emphasizing the scientific and flexible nature of supervision. The EU's regulatory style balances development and regulation, and it is expected to reshape the global digital development model through high-standard legislation and regulation. In terms of regulatory approach, the United States focuses on technology application and process supervision; The EU focuses on organizational management and risk management. In addition, there are also conflicts of ideas between the two sides in areas such as data governance and privacy protection. There are ** reports that the United States has warned the European Union that its proposed AI regulations will benefit those companies that have the resources to bear the cost of compliance and hurt those smaller ones, which is a concrete manifestation of the conflict of ideas between the two sides.
The world will work together to build an environment for the development of artificial intelligence
At the first AI Security Summit held in early November, representatives from the United States, the United Kingdom, the European Union, China, India and other parties discussed the risks and opportunities brought about by the rapid development of AI technology. Following the meeting, 28 countries and the European Union signed the Bletchley Declaration, pledging to work together to build trustworthy and responsible AI. The host country, the United Kingdom, also announced that the next AI Security Summit will be held in France in a year's time, and South Korea will co-host a mini-virtual summit in the next six months.
It is foreseeable that with the development and popularization of artificial intelligence, the regulation and regulation of artificial intelligence by the international community will become a global issue. The legislation and cooperation between Europe and the United States in the field of artificial intelligence have certain implications for other countries around the world. However, in the process of establishing and improving the international AI governance standard system, in addition to a few "leaders", there are also many participants. All countries in the world should actively exert their influence and jointly promote the construction of an equal, open and mutually beneficial environment for AI development.
*:Light**.