news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

On August 1, local time, the EU Artificial Intelligence Act officially came into effect. As the world's first comprehensive regulation on artificial intelligence, its introduction marks an important step forward in the standardized application of artificial intelligence.

Will China's artificial intelligence companies, which are catching up, be affected by this?

How to regulate?

The bill adopts a risk-based management model, dividing artificial intelligence systems into four categories: prohibited, high risk, limited risk, and minimum risk, and sets different compliance standards accordingly.

Among them, the definition and requirements of high-risk AI systems have become the focus of attention. According to the bill, self-driving cars, medical equipment, loan decision systems, education scoring and remote biometric systems are all high-risk AI systems, involving key areas such as medical health, public safety, and transportation. Therefore, the bill stipulates strict transparency obligations for high-risk AI systems.

At the same time, complex and widely used artificial intelligence systems will also be subject to new constraints. The bill requires that all AI-generated content (including images, audio or video, etc.) must be clearly labeled and can be detected as AI-generated content to address people's concerns about false information.

In addition, the bill prohibits the use of artificial intelligence systems that are considered to pose a clear threat to the basic rights of users. For companies that violate the law, the EU will impose a fine of up to 35 million euros or 7% of its global annual turnover, whichever is higher.

According to the plan, the relevant rules of the bill will be implemented in stages, mainly to give companies a certain transition period. Some rules will take effect six months or 12 months after the bill is passed, while most rules will take effect on August 2, 2026.

It is worth mentioning that this bill will not only apply within the EU, but its effectiveness will also extend to related companies operating outside the EU, and it has broad international influence.

What is the impact?

As the provisions of the bill are gradually implemented, the application of artificial intelligence may move towards a more standardized track.

Fang Yu, consultant of Beijing King & Wood MallesonsWhen interviewed by China News Service's National Affairs Express, he pointed out that the EU's "Artificial Intelligence Act" has maintained the same "high standards" from restrictive content to penalty provisions, has a benchmark significance, and will provide a reference framework for legislation in other regions and countries in the world.

At the same time, there are many opinions in the industry that compliance means rising costs, and the implementation of this bill will significantly increase the compliance costs of AI-related companies.

According to early forecasts, the bill could cause the EU economic losses of more than 30 billion euros over the next five years and increase compliance costs by nearly 17% in the field of artificial intelligence.

Peng Kai, partner of Beijing Dacheng (Shanghai) Law FirmIt is believed that the AI ​​Act has high requirements for transparency and personal information protection, which is good for consumers. However, given the strict regulatory requirements, participants in the AI ​​value chain need to invest more resources. Some companies may need to adjust their market entry strategies or even give up entering the EU market because they cannot afford the high compliance costs.

In Fang Yu's view, if Chinese artificial intelligence companies want to enter the EU market, the compliance costs and pressures they face will inevitably increase. In addition to compliance costs, he also emphasized that companies may face "value costs."

Fang Yu believes that for overseas companies, the aspects and impacts of this bill are complex and multifaceted. The bill is based on risk governance, but there are certain differences between the EU and China in the understanding and definition of "risk". Chinese overseas companies have to face not only institutional compliance, but also certain adjustments and balances in terms of values ​​and other cognitive aspects.

Hu Naying, Senior Director of AI Governance, AI Research Institute, China Academy of Information and Communications TechnologyHe said that after the law comes into effect, the challenges faced by Chinese overseas companies are, on the one hand, to adjust their own business and compliance systems internally to align with EU standards, and on the other hand, to balance the rules between different countries and jurisdictions externally.

Some people are concerned that the bill may inhibit technological innovation due to stricter supervision and constraints, especially putting greater pressure on start-ups and high-risk AI system suppliers. However, some analysts believe that as the compliance system improves, in the long run it will inspire companies to reduce compliance costs through technological innovation and promote the standardized development of the AI ​​industry.

Fang Yu said that the bill itself is just a tool, and different companies have different compliance standards, so it cannot be generalized. However, if too many companies or categories are included in the high-risk level, it may affect the innovation of companies to a certain extent.

Column Editor: Qin Hong Text Editor: Song Hui Title Image Source: Tuchong Image Editor: Xu Jiamin

Source: Author: National Express