project details

In RegTech4AI, we seek to pursue four mutually reinforcing Aims relating to AI regulation in the EU:

  1. Regulation: assess and improve the EU’s process for AI regulation
  2. Communication: make knowledge around AI regulation more accessible to SMEs and start-ups
  3. Competition: ensure fair competition in AI and mitigate platform power
  4. Enforcement: build new methods and metrics to help enforcement of AI regulation at scale

The focus on these four Aims in RegTech4AI will directly support the trustworthy and human-centric adoption of AI in the EU in line with long-established legal principles, like transparency, non-discrimination and protection from bias, fairness, privacy and data protection, and human agency. The novelty and strategic importance of RegTech4AI lies in helping implement AI-relevant laws like GDPR and the AI Act and making them work in practice. This has rarely been studied in previous work, but is a necessity for the implementation of those laws and for the effective protection of citizens in the age of AI – thereby bolstering trust in AI systems among citizens.

Overview of RegTech4AI, the research project.

Aim 1 aims to lay the foundation for the rest of the research and to study current flaws in the EU’s approach to regulating AI. The findings from Aim 1 will feed into Aims 2–4, which aim to build novel regulatory technologies (RegTech) for the areas of communication of AI laws to businesses, the mitigation of disproportionate market power in AI technologies, and the enforcement of AI laws in practice. Aims 2–4 reinforce each other because they all apply technical (and other) methods to AI regulation. For each of the four Aims, we will hire a PhD student with the relevant expertise in computer science or law, and pair law and CS students as “buddies” to ensure interdisciplinary dialogue. All PhD students are expected to extensively read laws and produce program code (e.g. for data analysis or building research prototypes). The research will be supported by a postdoc with expertise in interdisciplinary research, so as to involve relevant stakeholders, conduct surveys and interviews where required, and convene online workshops. Since we focus on theoretical and empirical research, we foresee very limited risks to the feasibility of this proposed research project.

RegTech4AI will primarily focus on the GDPR and the EU AI Act. The GDPR remains one of the most advanced AI-relevant laws. It is the blueprint for the AI Act as well as all other new tech laws in the EU. While the GDPR has a legal history of 40+ years, the AI Act is a brand-new type of law; it will therefore take years of clarification in the EU courts until it reaches the necessary clarity. By focusing mainly on these two laws, we aim to derive valuable lessons for the regulation of AI in the EU. Our fifth Aim cuts across all four aims and seeks to build crucial bridges between computer science (CS) and law, the lack of which creates a significant barrier to regulating AI. This fifth Aim is addressed by our focus on the four other Aims, as well as a range of supplementary activities.