Members of the European Parliament (EP) have reached a provisional political deal on the Artificial Intelligence (AI) Act: "We have a deal now in which all groups will have to support the compromise without the possibility of tabling alternative amendments", a Member of the European Parliament told Euractiv.

The AI Act is now expected to be put to vote in the Committee on the Internal Market and Consumer Protection on 11 May 2023, and then in the EP plenary in mid-June.

Some of the elements that have sparked intense discussions in the EP so far have been related to the inclusion of provisions on generative AI (including a requirement for companies that deploy generative AI tools to disclose the use of copyrighted material), and the types of AI applications that are to be banned, the classification of high-risk systems.

The deal in the EP comes two year after the European Commission issued its proposal for the AI Act which follows a risk-based approach by differentiating AI tools that create low, unacceptable, and high risk. The assessment of the high-risk classification will be based on the intended purpose of the AI tools and the product whose safety component is required to go under third-party conformity assessment. And while some AI tools may be considered to create a high risk to health safety or fundamental rights, they will not be banned from European Market. Instead, they must be placed in the market under EU's harmonization rules on risk management, transparency, and data governance. Additionally, the AI Act sets a mandatory code of conduct for high-risk AI tools requiring them to commit to environmental sustainability, accessibility for persons with disability, participation of stakeholders in designing AI tools, and diversity of development teams.