Talking past each other? Navigating discourse on ethical AI: Comparing the discourse on ethical AI policy by Big Tech companies and the European Commission

This study examines the convergence of the European Commission (EC) and Big Tech companies' (Google and Microsoft) discourse on 'ethical' AI through critical discourse analysis and the concept of hegemonic discourse. The paper answers the question to what extent there is a hegemonic discourse on ethical AI between EU policymakers and Big Tech companies and whether this is impacted by the prospect of legally binding legislation, considering the possible impact of the 2021 AI Act Proposal of the European Commission. This analysis is relevant at an inflection point where previous literature notes superficial convergence between the approaches of public and private actors, indicating policy consensus. The scope of analysis however is limited to non-legally binding regulation and lacks regional focus. In the EU, the advent of legally binding AI regulation with the 2021 AI Act (AIA) Proposal marks a critical juncture: with agreement on the AIA in December 2023, ethics standards become part of market entry requirements to the EU Single Market and the underlying differences in approaching Ethical AI will have important ramifications on policy preferences, compliance, enforcement and thought leadership in the domain more broadly. I find that the European discourse on 'ethical' AI by the EC and Big Tech companies such as Google and Microsoft is largely hegemonic and depoliticised in non-legally binding settings from 2018-2021 due to shared assumptions on 'ethical' AI and absence of significant underlying social and political conflict. It evolves to non-hegemonic and repoliticised discourse through dislocation by the prospect of legally binding regulation, which pushes actors to reveal their genuine policy preferences that bear political and social conflictuality whilst both actor types take an instrumental approach to ethics.