The European Union has earlier issued a bill on the safe development of artificial intelligenceReaching an Agreement, a reference standard covering a wide range of projects will be established to avoid potential social risks caused by artificial intelligence.
Dragoş Tudorache, a Romanian MP who co-sponsored the bill, said it would become a benchmark for many jurisdictions.
The proposed regulations will cover how companies develop and deploy artificial intelligence, as well as its application principles in education, employment, healthcare and other fields. They will also define the potential risks of artificial intelligence at four levels: minimum, limited, high and prohibited.
Among them, circumventing the user's wishes, collecting data from protected social groups, or even tracking and recording personal information through real-time biometric features such as facial recognition at any time will be listed as prohibited. As for technologies used in critical infrastructure, education, legal or judicial-related businesses, and employee recruitment-related technologies, they will be listed as high-risk. Application services provided by companies such as ChatGPT, Bard, and Copilot are currently listed as limited risk.
The EU has already proposed relevant management measures for artificial intelligence technology in the past. For example, in 2018, it proposed an artificial intelligence strategy and artificial intelligence coordination plan for Europe, and in 2019, it proposedEthical Guidelines for the Application of Artificial Intelligence Technology, was established in April this yearSpecial Task Force, helping countries deal with data protection measures behind artificial intelligence technology.
Prior to this, the European Union had already proposed the General Data Protection Regulation (GDPR) to protect the privacy of European residents, requiring all services to obtain user consent before collecting necessary data. In the current data-heavy application of artificial intelligence technology, the EU hopes to strike a balance between data collection and technological application, preventing companies from arbitrarily collecting user privacy under the pretext of using AI technology, while also preventing technological development from being restricted by regulations.
In the draft artificial intelligence regulation proposed by the European Union, it is emphasized that artificial intelligence must be a tool to serve humanity, and its ultimate goal is to increase human welfare, rather than developing artificial intelligence technology itself as the main purpose. Therefore, the use of artificial intelligence in the market must be people-oriented and developed with trustworthy technology. At the same time, it must be provided for use while protecting basic human rights and complying with laws and regulations. It must also comply with the principles of fairness and equality and cannot restrict or hinder the development of other technologies.
