The AI Act is so last year. Or: What is the Convention on Artificial Intelligence and why is nobody talking about it?

European AI Governance is moving fast. In this talk we aim to untangle, analyse, and compare the two main regulatory proposals on the table. The AI Act (AIA) and the Convention on Artificial Intelligence (CAI). Having followed both in the past years, we hope to bring more clarity on these related sets of rules, which will shape how automated decisions are regulated in the coming decades.

The AIA is an EU  product regulation and applies in all 27 Member States (like the GDPR). The CAI is a proposed treaty of the Council of Europe (which is not an EU institution and has 47 member states). So the CAI is more akin to the European Convention on Human Rights. The AIA was proposed in 2021, the CAI in 2022, and the negotiators are aiming to make both complementary. While the AIA aims to protect health, safety, and fundamental rights, the CAI aims at the protection of human rights, democracy, and the rule of law. Both propose very specific but slightly different bans and risk management systems for high risk AI applications. The AIA is likely to come into force earlier and so far has received more attention but the CAI is no less relevant as it could potentially have an even wider reach.

In this talk, we want to provide a comparative overview of both proposals and and raise some critical questions about them:

• What is at stake for the lives and rights of millions of people who will be affected by these rules? What changes will they mean for your life? What are potential loopholes in both proposals?
• Why are both proposals so far receiving fairly little attention and why are they mostly discussed behind closed doors?
• What are the consequences of corporate lobbyists having far more resources to accompany EU policy processes like the AIA? Why have civil society observers been banned from the negotiations of the CAI?
• What effects will these rules have beyond Europe? And looking at what regulations have already been proposed in other jurisdictions since they were presented: what effects are we already seeing?
• When and how can we still have an influence on these processes? In which ways can people participate in the sometimes opaque and legalistic world of European tech regulation?