G7 countries agreed on principles for regulating artificial intelligence (AI)

On Monday, the G7 countries released a joint statement agreeing on “regulatory guidelines” for artificial intelligence (AI). The G7 calls on other states to pay attention to the possible risks associated with the rapid development of AI, and also formulate an approach to regulating this industry. The EU has already welcomed the G7 statement, which is generally consistent with the approach of the European Union itself to the problem of regulating AI.

On October 30, at the G7 summit held in Hiroshima, Japan, a joint statement was published (.pdf) G7 countries on the basic principles of regulation of the AI ​​industry. The 11 statements do not provide detailed guidance to companies or organizations, but rather outline key issues and risks – with suggestions on how they can be avoided or minimized.

“This list of principles is not final and may be supplemented based on existing OECD principles on AI in response to recent developments in the modern AI industry and designed to both take advantage of all the benefits of these technologies and prevent possible risks. These principles can apply to all parties working in the field of AI—whether it is the design, development, promotion, or practical application of advanced AI systems,” the statement said.

The first principle in the G7 statement is the need for timely risk assessment and testing of new AI systems – not only by the developers themselves, but also with the help of special working and/or expert groups. The corollary of this principle is that companies and organizations involved in the development and implementation of AI should be encouraged to use independent experts to assess the potential risks of AI, who would also help developers create mechanisms to solve problems that arise during the testing stage.

The G7 also called on AI developers to be as transparent as possible when working on new AI systems, emphasizing that the disclosure process itself should be as simple and accessible as possible.

Another guiding principle for regulating the AI ​​industry, the G7 considers “the need for timely and responsible informing of specialized organizations, the expert community, government agencies and the general public about incidents that arise during the development or testing of new AI systems.”

The responsibility of developers, according to the G7, should also be “to establish security and control mechanisms – including physical security, cybersecurity and protection against insider threats throughout the development and implementation cycle of AI systems.” The G7 countries also believe that priority areas in the development of modern AI should be the most important problems of the modern world, including the fight against climate change, health/medicine, education, etc.

Another important guiding principle is the work to preserve and protect intellectual property in the field of AI, ensuring high-quality storage and transmission of information, legal and technological measures to prevent the abuse of these developments. “We encourage all stakeholders, companies, organizations, governments to take these principles into account when developing their own regulatory standards in the field of AI,” the G7 countries noted.

A few hours after the publication of his statement commented European Commission. She welcomed the development of principles for the responsible and safe development of AI, saying that their spirit was consistent with the EU’s own vision in this area. European Commission President Ursula von der Leyen noted that “the potential benefits of AI for citizens and the economy are enormous. However, accelerating AI capabilities also brings challenges. The EU is already at the forefront of developing AI regulations, contributing to the global process. We are pleased to welcome the G7 International AI Guidelines, which also reflect the EU’s position on responsible AI development.”

Evgeny Khvostik