Artificial intelligence (AI) offers great promise in terms of societal and economic benefits. But it also raises numerous ethical and human rights concerns as the technology might inherit and exacerbate flaws that are already present in society. There have been numerous calls for the regulation and legislation of AI, most prominently the EU’s 2021 proposal for an AI Act. A recent Computer Law & Security Review publication explores what an AI governance body could look like on the European level.
AI raises concerns of bias and discrimination, privacy and security, economic and other inequalities, political and power imbalances, and a restructuring of human interaction, thoughts and lives. The regulatory challenge is that of harnessing the benefits the technology can bring while mitigating the risks. According to the authors of a recent paper, successful AI regulation is likely to require the creation of an organisation that will act as a gatekeeper and maybe as a regulator. They have reviewed existing institutions and now propose an outline for an EU AI Agency.
“What we have attempted to answer is the question of how a European Agency for AI should be designed to promote human flourishing. Stepping outside the academic world to provide a practical contribution to an evidence-informed policy debate and development, I think I speak for all the authors when I say that our proposal will have both theoretical and practical impact, but it is difficult to predict the exact outcome of political negotiation,” says Bernd Stahl, Professor at De Montfort University in the UK and one of the authors of the paper.
The paper discusses how the construction of an EU AI Agency differs from the proposal for an EU AI Board that was proposed in the EU AI Act and makes recommendations for how it could be improved. This includes increasing its independence from the European Commission, enhancing diversity with permanent representation from a wider range of stakeholders, expanding its mission to include more tasks and functions, and adopting operational principles.
“Regulatory bodies for AI will likely come into existence in the very near future and there is much work to be done. Our goal is to inform the political discussions that will precede their creation on the European level. On the national level, we foresee a need for similar considerations also taking into account different cultural settings and political traditions. And there is also a need for global coordination,” Bernd Stahl concludes.
By Anna Holm
Stahl BC, Rodrigues R, Santiago N & Macnish K, A European Agency for Artificial Intelligence: Protecting fundamental rights and ethical values, Computer Law & Security Review, July 2022: 45; 105661