Google in Brussels: A technology giant in the co-construction of a European artificial intelligence policy

Artificial intelligence (AI) is one of the priorities of the European Commission, as highlighted by the recent increase of its dedicated budget and the multiplication of communications and white papers produced by its services. These efforts are part of a will to ensure “Europe's digital sovereignty” currently challenged by the Chinese and US-based tech giants that dominate the sector. Most of the latter, such as Google or Facebook have increasingly invested in lobbying the European institutions over the last decade. This paper specifically focuses on Google . Created in 1998, and famous for its search engine, Google has expanded its business model with multiple services and products based on AI. Indeed, since 2009 it has acquired 30 AI start-up making it the first AI start-up purchaser company - in the world. The theoretical premiss of this paper is that fabrication of public policies is a co-construction between the institutions and the private actors (Laurens, 2019). Thus, the co-production of a European artificial intelligence policy that is our central question: To what extent has this public policy been co-produced with Google? To answer this, three sub-questions must be tackled. First, we need to question the modalities of entry and the evolution of Google within the Brussels bubble thought as a field (Bourdieu, 2015), the field of the Eurocracy (Georgakakis, 2014). Secondly, we must ask ourselves how Google is considered by the services of the European Commission and more particularly DG CONNECT. Thirdly, we will see how Google interacts with the other stakeholders in this field. From a methodological standpoint, this communication relies on several types of sources. First, we used the documents produced by the European institutions, NGOs, Google, the Belgian administration, and press articles as well. Then we conducted a prosopography based on the list of members of the High-Level expert group on artificial intelligence, set up in 2018. This group, to which Google belongs, oversees the production of ethics guidelines for trustworthy artificial intelligence and delivers policy and investment recommendations for trustworthy artificial intelligence. Finally, I based my research on ethnographic sources: tree semi-structured interviews (conducted between December 2019 and March 2020) and a non-participatory observation of the Master of Digital event organised by Digital Europe (One of the biggest group interest focused on digital policy and it is based in Brussels) on the 6th of February 2020, during which Google is a sponsor and participated in a panel on artificial intelligence.