The European Parliament’s Civil Liberties and Internal Market committees this week approved the text of the proposed EU law on artificial intelligence by a large majority. The law, which has been in the works since 2021, aims to identify and address the threats posed by artificial intelligence to the public, particularly in the areas of copyright and privacy. If passed, the law could be the world’s first rules regulating the use of AI.
The EU AI Act will assign risk factors to any AI product or platform. These categories range from unacceptable to minimal. The benchmark will enable developers and engineers to consider how to bring their AI product to the European market. The new law will ban any AI systems it deems “unacceptable” in the EU, such as those that pose a threat to public health, safety or fundamental rights.
For regulatory purposes, legislators have divided the “General Purpose AI” (GPAI) categories into levels. Most of the reporting and regulatory burden related to GPAI will fall on those market participants who integrate these systems into an application that risks posing a threat to public health, safety or fundamental rights. High-risk AI providers face more stringent regulations. This applies in particular to risk management, data management, technical documentation and record keeping.
The vote in the parliamentary committee is the first step in a difficult legislative process. If the EU AI Act is passed in the EU plenary next month as planned, the bloc will send the Act to the European Council for negotiation. At that point, all EU member states will have the opportunity to express their views on the final form of the law.
