Debate: Is the EU doing enough on Artificial Intelligence?

Artificial Intelligence technologies, especially in the current coronavirus pandemic crisis, are called to transform our societies. However, its use raises questions of all kinds. How should they be regulated? What are the ethical implications? What effects will they have on democracies? At the EU level, is the European Union doing enough to enforce its AI strategy and position itself in the global race, where China and the United States are key actors? How can we ensure a human-centered AI?

On April 23rd, IDEES magazine, the Centre for Contemporary Studies (CETC) and the Delegation of the Government of Catalonia to the European Union hosted an online debate entitled Is the EU doing enough on AI?, which featured panelists who have expertise in different areas of AI and whose diverse backgrounds contributed to shedding light on the controversial issues and challenges that Europe is facing today regarding AI. The discussion was part of a series of events organized to feature the latest IDEES magazine special issue on Artificial Intelligence. The debate was initially hosted in Brussels, but due to the pandemic it was hold online on Youtube streaming.

The debate was introduced by Meritxell Serret, representative of the Government of Catalonia to the European Union, and Pere Almeda, director of the Centre for Contemporary Studies, and was moderated by Carles Sierra, director of the Artificial Intelligence Research Institute (IIIA). The participants were Lorena Jaume-Palasí, Executive Director at the Ethical Tech Society, Michela Milano, Vice President at EurAI, Andrea Renda, Head of Global Governance, Regulation, Innovation and Digital Economy at CEPS and Eric Badiqué, Adviser for Artificial Intelligence at European Commission.

The role of the EU

The debate covered many aspects regarding the alignment of AI with European values, the regulatory challenges that the nature of AI brings about, the role of the EU in the coordination of initiatives and its presence in the global scene, the present and future of research and the public-private relationship. All of the participants agreed that, while the EU seems to be speeding up its initiative in the field of AI, there is still a lot of work that needs to be done to ensure its successful and safe development.

Lorena Jaume-Palasí argued how, as happens with other kinds of infrastructures, the criteria that guide the creation of AI are not individual-based but collectivity-based. AI does not rest upon assumptions about individuals but upon an overview of societies and groups, and this collides with the very nature of legal regulation in western societies, which is focused on the individual. Jaume-Palasí argued that regulation in areas such as AI has to be based on societal values, such as social cohesion or sustainability. The EU has already taken this approach in other fields, but not in AI, an area in which such a perspective is crucial. Jaume-Palasí also warned that today’s decisions will have an impact on democracy, especially in the current pandemic situation. “A big part of the democratic architecture that we have today was created in war and epidemic situations. The things we change now will remain and shape the evolution of democracies. We need concerted action from different disciplines”.

Michela Milano provided an overview from the point of view of research in the field of AI and analyzed the weak aspects of AI research in Europe, to which she responded by proposing some solutions. The fragmentation of initiatives that she detected could be countered with the creation of collaboration and the strengthening of the research ecosystem; the lack of alignment between basic research and applied industrial research could be addressed by ensuring the connection between both areas; the inability for retaining talent within the EU could be tackled by providing more funding and reshaping research programs and, lastly, the limited access to AI for low-tech users could be overcome by making it more accessible. According to Milano, “research in AI is very important in the long term and the EU could use this technology and make a difference by solving sustainable development issues in the framework of the Green Deal”.

Collective-based legislation

Andrea Renda agreed on the need for a shift from individual-based legislation to a collective-based one and one that could take into consideration the change in human relations that technological mediation brings about. He also placed emphasis on the need to think about the future when designing a regulatory framework and on the idea that AI has to be human-centered and dealt with more as a means than as an end. “The question is not what the EU can do in AI, but what can AI do to help the EU to achieve its goals. For example: how can AI help with the SDGs?”. According to Renda, the EU has succeeded in implementing an ethical framework by adopting rules that are extraterritorial, risk-based and that ensure trustworthy AI. The important question for Renda is whether the EU can sustain such a legislation and still be competitive, to which he answers that the EU needs to lead the field of research and coordinate an international alliance of human-centric AI.

Eric Badiqué emphasized the idea that AI needs to serve and empower humans, and argued that the EU has done a lot on AI, especially when it comes to the discussion on ethics. However, he recognized there is still a lot of work to do. The EU is leading the conversation on AI at a global level but still needs to perfect aspects related to funding and research, and the current Covid-19 crisis is only making it more difficult, as this highly depends on the EU budget. According to Badiqué, the role of the EU in the world is crucial because it could establish a sort of third way between the US and China and open a track to work towards a broader consensus between like-minded countries around the world. He concluded his intervention by making some observations about the lessons Covid-19 has taught us so far about digitalization, trust and data and raising some thought-provoking questions for further reflection, such as “can we expect that the COVIDー19 crisis will lead to a higher awareness of privacy-related issues globally? Or will it go the other way?”.

Final debate

The debate ended with a round of questions from the audience, which they sent via the hashtag #EUonAI on Twitter and also via the live chat of the broadcast. The panelists covered some issues that had not been tackled and clarified those that had been more controversial. The main lines of the final debate where about how to implement risk governance, how to coordinate the public-private interaction, the need to ensure an interdisciplinary approach to AI, how to ensure a proportionate decrease in privacy in times of Covid-19, how to adjust to Brexit, how to regulate foreign apps entering the European market, how to retain talent and ensure diversity and, lastly, the geopolitical implications of AI and the impact of the 5G struggle on democratic principles and fundamental rights.