The use of Artificial Intelligence (AI) in health is said to hold great potential for improving healthcare worldwide. AI systems are deployed to recognize patterns in enormous datasets, predict a certain outcome, and take action accordingly. The use of health-related AI could improve the quality, effectivity, efficiency, and accessibility of healthcare. To illustrate, a large-scale study conducted in German hospitals taking part in the national breast cancer screening program, revealed that an AI recommender system for breast-cancer screening significantly improved the screening accuracy of radiologists.

At the same time, AI may deepen existing global health disparities. The development of medical AI requires large amounts of high-quality medical data. This data​ on European patients​ is scarce because of thorough regulation in EU Member States. Big tech companies (e.g., Google, Facebook, Microsoft) are therefore increasingly collecting these data in Low- and Middle-Income Countries (LMICs), taking advantage of the limited legal and regulatory frameworks surrounding personal data protection and product regulation. The products developed with these data are then marketed in the Global North​, which is an ​emerging practice​​ described as “data colonialism”.

​​Data colonialism​​ is however not a new phenomenon: there are many examples of medical researchers escaping restrictive regulatory regimes ​in some parts of the world ​by exporting otherwise prohibited research practices to LMICs – referred to by the European Commission as “ethics dumping”. In this way, researchers located in an EU Member State do not have to comply with the strict EU framework for clinical trials. ​​In many cases, burdens and risks are imposed on people without sharing the benefits, such as the profitable export of human tissue materials​ from LMICs to develop medicines for countries in the Global North.

​​​​​

​​We argue that the​​​ rapid emergence of research and development of “data-hungry” AI – governed by strict EU rules – may give rise to a new reality of “digital ethics dumping”. Research already suggests that the adoption of strict rules for data protection in the EU seems to affect the exporting of clinical trials to LMICs. The new regulation for AI products proposed by the European Commission in April 2021 may amplify this effect.

​​​​​

​​​In this paper we explore the parallels between the experience of outsourcing drug development, specifically clinical trials, to LMICs and the development of health AI.​​

Stakeholder consultation is a requirement for virtually all policy initiatives of the European Commission and a key tool in the EU’s pursuit of ‘better regulation’. Its increasing use reflects its (perceived) capacity to enhance legitimacy, address democratic deficits and contribute to evidence-based policy-making (Bunea & Chrisp, 2022; Kohler-Koch & Quittkat, 2013). However, consultation also has the potential to produce or exacerbate bias, offer additional lobbying opportunities, and broadly ‘stack the deck’ in favour of particular interests (Braun & Busuioc, 2020; Bunea, 2017; Quittkat & Kotzian, 2011). This paper takes a discursive institutionalist approach to the latter possibility, exploring the way in which stakeholder consultation empowers actors and ideas.

The paper focuses on the design and execution of consultation strategies, and how this might privilege particular actors and ideas. Drawing from the literature on stakeholder engagement and the EU consultation regime, as well as that on interest groups and the EU lobbying regime, the paper explores how consultation design – the instruments used, actors targeted, input solicited – might give voice to particular constituencies and establish particular ideas about the policy problem, solution or context. We take as our case study the ongoing revision of the General Pharmaceutical Legislation (GPL), a key piece of EU health legislation. We review the consultation strategy for the GPL revision, before mapping the activities undertaken and the instruments used. The paper then offers a thematic analysis of how ideas about process (e.g. what constitutes appropriate evidence) and substance (e.g. what drives unequal access to medicines) are embodied in the consultation design and tools, and an analysis of participation patterns across stakeholder groups, aimed at providing a nuanced picture of engagement that goes beyond quantitative description to include origins and depth of interaction.

In its methods and approach, the paper offers some innovation. The majority of the existing research on consultation and interest groups focuses on the interaction between institutions and actors. We add to this an exploration of ideas, broadening our understanding of how stakeholder participation impacts policy-making. The paper presents an issue-specific, qualitative case study, complementing and testing some of the findings of the large-n, quantitative approaches that dominate the literature. Moreover, in drawing on data gathered via freedom of information requests, and reflecting on this methodological tool, we seek to contribute to advancing the study of participation in (EU) public policy-making.

Unavailable and unaffordable medicines (i.e. vaccines and therapeutics) increasingly threaten health systems, and political and economic stability. Persistent high medicines prices (e.g. orphan/cancer medicines) and few new medicines (e.g. antibiotics, vaccines) lead to a vicious cycle of hoarding, inefficiencies, and risk to human health worldwide. These issues arise from the current model of drug research and development (R&D), an area in which the European Union (EU) has amassed significant power, and with it, the potential to shape medicines access globally. Now, calls from civil society and democratic representatives implore the EU to improve its approach to global access to medicines.

A major challenge towards a well-defined role for the EU in access to medicines in low- and middle-income countries (LMICs) is that there is no clear framework about which legal and bioethical principles, and EU values should guide EU decision-making in the field of pharmaceuticals. The competing, and sometimes contradictory, nature of EU pharmaceutical policy, together with the overlapping legal and regulatory regimes (national/EU/international) that influence medicines development, compound this challenge. Failure to clarify the EU’s responsibilities towards access to medicines in LMICs is likely to have consequences for the EU’s internal politics, public health, security, soft power, and trade relations, far beyond those seen during the Covid-19 pandemic.

Through a fundamental human rights lens, this project addresses the question: How should the EU conceptualize its responsibilities towards pharmaceuticals in LMICs?

This project builds a conceptual framework for ‘EU-global access to medicines’ using doctrinal legal analysis based on primary and secondary sources of international, EU, and Member States' constitutional law (codified and authoritative interpretations of fundamental and human rights and bioethics). This research brings together the EU’s constitutional responsibilities and core values related to health (e.g. public health, development/international cooperation, humanitarian aid, solidarity, human rights protection in Arts 168, 208, 214, 222 TFEU, and Arts 6 and 21 TEU). This research also considers how the EU's legal principles and values interact with other legal responsibilities in EU MS’ constitutional law and international law (such as the International Covanent on Economic, Social, and Cultural Rights; International Covanent on Civil and Political Rights; Convention on the Rights of People with Disabilities), which enshrine a patchwork of ‘positive’ state duties towards health, medicines, and international cooperation. Ultimately, this project identifies the legal responsibilities of the EU towards global access to medicines.