THE LANDMARK EUROPEAN AI ACT IS MOVING DOWN THE LEGISLATIVE PROCESS
The landmark European AI Act is making its way through the EU’s ordinary legislative procedure. Once adopted and in force, the Act will impact organisations from a range of sectors that develop, deploy or use AI technologies. Based on the “Brussels Effect”, the AI Act will also affect organisations outside the EU – by covering AI systems placed on the EU market from third countries.
Due to the major societal and business relevance of the draft law and the extensive technical and political debates, the legislative procedure is particularly lengthy and complex. This Acquis EU Law & Policy update takes stock of the state-of-play and provides an outlook on what to expect next.
Nearly one year ago, in April 2021, the European Commission presented its much-anticipated proposal for horizontal legislation on artificial intelligence (AI). The “Proposal for a Regulation of the European Parliament and of the Council laying down harmonised rules on artificial intelligence” (“Artificial Intelligence Act” or “AI Act”) is expected to have far-reaching effects on organisations developing, deploying or using AI technologies. The AI Act aims to provide a legal framework to ensure the safety and fundamental rights of people and businesses in relation to AI systems placed on the EU market. At the same time, the rules seek to increase AI uptake and facilitate investments and innovation.
Key Elements of the AI Act & Debate
The draft AI Act regulates the specific utilisation of AI systems and associated risks. It includes a ban on specific AI systems posing “unacceptable risk”. The draft Act introduces strict requirements and obligations on a larger number of “high-risk” AI systems. A dedicated EU regulatory authority will enforce compliance, and the framework foresees significant fines for infringements. The legal framework will impact a range of economic sectors, including health, energy, transport, agriculture, tourism and cyber security. Importantly, it will also have “extraterritorial effect”, covering AI systems placed on the EU market from third countries.
The main sticking points in the legislative debate include the scope of the definition of AI, banning facial recognition, the breadth of the category of high-risk AI, banning of social scoring from private companies, and whether to require additional third-party checks beyond self-assessments for most AI uses.
Procedure & State-of-Play
A large number of organisations and individuals are seeking to exert influence on the legislative process by engaging with the co-legislators, namely the 27 Member States in the Council and the European Parliament. On the part of the European Parliament, the Internal Market (“IMCO”) and Civil Liberties (“LIBE”) Committees are leading the negotiations, with 5 more committees involved, each with a special focus on certain provisions. This makes for an unusually large number of committees and Members of the European Parliament (“MEPs”) seized of the file, which creates additional complexity. Likewise, within the Council, negotiations to find a common position amongst the Member States are in full swing.
Related EU Upcoming Legislation
The AI Act is one prominent file amongst a number of key recent EU pieces of legislation in the digital field. This includes the Digital Services and Digital Markets Act, the Data Governance Act, and the Data Act. Similarly as with the landmark EU General Data Protection Regulation (“GDPR”), it is widely expected that the AI Act will also have an effect in terms of shaping future AI regulation in non-EU countries.
EU-US Cooperation
A related element important to follow is the relaunch of transatlantic cooperation on trade and tech policy in the framework of the EU-US Trade and Technology Council (“TTC”) since June 2021. One focus of cooperation will lie in jointly regulating big tech and standard-setting, as well as involving other third countries. Building upon the work conducted by the ten dedicated working groups, the second TTC summit will take place in France in May 2022. Under the Biden Administration, the US has taken on a more proactive stance with regard to AI regulation, and like the EU favours a risk-based approach (Statement on AI, EU-US TTC Inaugural Joint Statement).
Timeline
Parliament’s plenary vote on the AI Act is currently scheduled for November 2022, although a delay should be accounted for. Following the publication of the partial compromise text by the preceding Slovenian Council Presidency, the current French Council Presidency is now pushing the file forward. Due to the technical and political complexity of the file and the major impact it is expected to have on business and society, the negotiations are expected to be lengthy. MEPs involved have indicated that discussions could last for two more years.
Following the adoption of the final text of the Regulation by the Parliament and Council, the rules will become directly applicable in all EU Member States.
Support by Acquis EU Law & Policy
Should you require assistance in assessing the impact of the European AI Act or other EU existing or upcoming legislation in the digital field on your organisation, please do not hesitate to contact Acquis EU Law & Policy. Our team will be glad to provide support and advice with regard to the options to have your interests represented.
***