German and English Sociotechnical Imaginaries in the AI Debate

The next speaker in this IAMCR 2023 session is Vanessa Richter, whose interest is in the shaping of AI debates and trajectories on Twitter. Imaginaries of AI are still evolving, and involve a diverse set of stakeholders: industry, governments, NGOs, academia, and the media. This project examined the accounts engaging in the debate on Twitter, and classified these into a number of different stakeholder categories; on Twitter, much of the debate appeared to be driven by AI experts (and the same is likely true on WeChat as well, as other research shows).

The present project builds on the idea of sociotechnical imaginaries to examine the public imaginaries about AI that are emerging from such discourse: publicly constructed visions of (un)desirable socio-technical futures. The project advances beyond a simple pro/con binary division, and instead maps the AI stakeholder environment using a more diverse set of categories. It gathered relevant discussion on Twitter in German and English, using relevant hashtags, and performed a social network analysis of retweet networks as well as a coding of the top 100 accounts per language and year, for 2012 to 2021.

German and English discourses had a very different mix of stakeholders, and overall those mixes also shifted substantially over the years. In Germany, the involvement of research, government, and NGO accounts grew over time, gradually replacing media accounts as the largest category; in English-language discussion, industry and (more recently) tech influencers came to dominate the discourse very substantially, replacing media and celebrity accounts. In the German discussion, there was also a growth in national and sutra-national accounts (partly driven by a greater discussion about AI regulation in the EU), while in English more specific individual and industry accounts came to dominate.

This shows some very distinct differences in AI discourse between these countries and regions, which matches what we know about the dominant themes of these national and supranational-national discourses, and points to the emergence of diverging AI imaginaries as well.