The European Council has officially approvedEU Artificial Intelligence DirectiveIt has officially come into effect in August and will use a unified artificial intelligence risk framework to formulate the responsibilities and obligations of artificial intelligence developers and users, and the Office of Artificial Intelligence under the European Commission will be responsible for promoting and implementing related business.
Under the EU Artificial Intelligence Directive, any AI deemed to have a significant impact on operational cognition, social credit rating, or possessing predictive law enforcement capabilities will be prohibited from use within the EU. AI designed for specific industries and uses, such as vehicle safety, healthcare, aviation, shipping, education, law enforcement, or the introduction of AI technology into specific products, will be subject to additional regulations to ensure safe use. AI technologies used for data analysis and that do not directly interact with humans will be categorized as relatively low-risk AI, and the regulations will be more flexible.
The EU Artificial Intelligence Directive will strictly prohibit artificial intelligence application functions and services that pose obvious risks to fundamental rights. For example, application services that may involve the processing of personal biometric information will be prohibited.
In addition, artificial intelligence technologies used in education, healthcare, law, border control, elections, etc. will be classified as high-risk projects and must comply with extremely strict requirements and regulations. The EU will also formulate relevant regulations for artificial intelligence technologies such as those used to filter spam, which means that all artificial intelligence technologies within the EU must comply with relevant regulations.
In order to enable corporate operations to comply with the relevant provisions of the EU Artificial Intelligence Directive more quickly, the European Commission is currently working on developing a general code of conduct for artificial intelligence, which is expected to be completed before April 2025.
After the EU Artificial Intelligence Act is officially implemented, the EU will also provide relevant adjustment periods. For example, artificial intelligence with the highest risk and that cannot be used within the EU will have an adjustment period of about half a year, while artificial intelligence of a general nature will have an adjustment period of one year. The relevant implementation clauses are expected to be officially implemented two years after the public announcement, and the additional regulations for the highest-risk artificial intelligence will be implemented within three to four years.
