The EU plans to regulate A.I.

In March 2023,  Gian Volpicelli wrote an article for Politico.eu saying that the Artificial Intelligence Act is a new piece of legislation that the European Union (EU) plans to use to control artificial intelligence (AI). The goal of this law is to guarantee that AI technology created and deployed in the EU conforms to moral and security requirements.

The new law's goal to develop various risk categories depending on the possible effects of AI systems on society is one of its most salient elements. The regulatory obligations will be increasingly onerous as the risk level rises. While some lower-risk systems will have more flexible restrictions, the most high-risk AI systems will be subject to obligatory pre-market conformity evaluations.

The rule will also forbid other uses of AI, such as those that use subliminal methods to control people or prey on their physical or mental weaknesses. Also, it would restrict the use of AI for monitoring, with some exceptions made for national security and other public reasons.

Regardless of where they are headquartered, all businesses operating within the EU will be subject to the rule, as would those operating outside the EU if they provide AI systems or services within the EU. As a result, organizations like OpenAI, the creator of the AI model upon which ChatGPT is built, will likewise be governed by the rule if they choose to conduct business in the EU.

The EU's initiative to govern AI is a part of a larger push to establish itself as a pioneer in the creation of moral and reliable AI. While some have applauded the law as a positive move, others have expressed worry that it may hinder innovation and disadvantage European businesses in global markets.

The Artificial Intelligence Act is an ambitious attempt to strike a balance between the need to safeguard people and society from possible damage and the potential advantages of AI. It will be interesting to watch how successful it is in attaining these objectives and how it is put into reality.