Governing AI's Intermediaries: Model Marketplaces, Platforms and Supply Chains

Like much of the digital economy before it, AI is consolidating into a platform business model. Intermediaries in this space are appearing around every corner to provide data, computing and storage, to host open-source models, to train, tweak and provide ethical guarantees, and to shape and control the interaction between AI and the devices we have in our pockets. At the same time, discussions around the governance of AI are jet-setting into the fanciful. What if we focused less on controlling a mythical runaway autonomous system and more on the intervention points on the ground, scrutinizing the ways that lead businesses within algorithmic supply chains are shaping interactions for their own benefit and to extract their own value?

In this lecture, Professor Michael Veale walks through aspects of the emerging business models of AI, the ways they are similar to challenges that have come before and the ways they might truly present new challenges. For example, under what conditions should a service hosting a machine learning model, like Hugging Face or GitHub, remove a model? How is governance on those platforms happening and changing today? Do sensible points to analyze and intervene in AI supply chains exist — and could they?

This talk draws in particular on four recent pieces of work in this space that Professor Veale has conducted with colleagues from Hong Kong University of Science and Technology (HKUST), WZB Berlin Social Science Center and University of Cambridge.

This event was hosted by Northeastern Law‘s Center for Law, Information and Creativity (CLIC) on April 2, 2024.