The European Artificial Intelligence Act (AIA): How it started, how it's going, and what's at stake.

Noah Schöppl

The GDPR has set a global standard for data protection. The AIA is set to do the same for AI. Given its potentially momentous impact, why is there so little attention on the ongoing negotiations in Brussels? In this talk we'll explore the EU's policy package on AI, its potential loopholes, and what's at stake for fundamental rights and democracy.

As controversial AI applications such as facial recognition, social scoring and automated decision making systems are proliferating, what is at stake for fundamental rights and democracy? This talk provides a deep dive into the AI policy process of the EU: starting with the Trustworthy AI Guidelines by the High Level Expert Group on AI in 2019, to the Whitepaper on the EU Approach to AI in 2020, culminating in the proposal for a European Artificial Intelligence Act (AIA) in 2021. This proposal encompasses limited bans for the most risky AI practices and is based on a regulatory risk pyramid, ranging from low to high risk AI applications with respectively increasing restrictions and requirements.

Some of the core controversies around the AIA will be considered, such as: Which AI practices should be categorized as low, medium, high or unacceptable risk? What restrictions should follow for each category? Should biometric recognition be banned and are the proposed exceptions justified? Yet, it will also entertain fundamental questions such as: is a risk based approach the right one and how broadly should AI be defined and scoped? Potential loopholes for national security, border control, general purpose AI, transport and R&D will be discussed critically. In this context, it will also analyze the connections with related policies such as the Digital Services Act and the Political Advertisement Regulation.

With the EU Single Market being the biggest market worldwide, the legislative package on AI is sure to have an effect on people, companies, and jurisdictions worldwide, a phenomenon termed “the Brussels Effect”. Thus, as with GDPR, the decisions on the AIA, mostly made behind closed EU doors, are likely to have a historic impact. The AIA is a momentous legislation, yet, so far there is little public discussion about its particular threats and opportunities. In order to make sure the AIA is protecting rather than threatening civil rights, the rule of law, and democracy, there is dire need for changes in the further legislative process of the AIA.

This talk will present the research of ALLAI, an independent organisation that was founded by 3 members of the EU High Level Expert Group on AI, Catelijne Muller, Virginia Dignum and Aimee van Wynsberghe. It is committed to drive the discussion around the AIA also outside the "Brussels bubble". 

Social Innovator and Policy Advisor