B  3  / Algorithms and Centralization

Increasingly, proposals to solve society’s hardest problems seem to carry the same tune: fire up a datacenter somewhere, let a fancy AI train itself, deploy the AI — that will solve the problem. But, there are many reasons to be cautious about the world-saving potential of algorithms. This seminar will consider one of those reasons — a crucial yet underappreciated feature of our new algorithmic world: the centralization of production and allocation decisions. Almost a century ago, economists hashed out a heated debate regarding whether, purely computationally, it would be feasible to have social decisions planned primarily by state-run algorithms (part 1 of the seminar). Many classes of algorithms such as the blockchain, recommendation systems, and generative systems, seem to march toward this dream. However, the algorithms built in the last 90 years are controlled not by a single public actor but by a range of private corporations, decentralized autonomous organizations (DAOs) and a patchwork of governmentrun entities (part 2 of the seminar). How can these algorithms maintain democratic legitimacy? What does this new form of centralization mean for democracy? This seminar will feature short lectures, seminar, coding demos, and creation of art.

Seminarleitung: Andreas Alexander Haupt und Zoë Hitzig, PhD