Economists generally see AI as a production technology, or input into production. But maybe AI is actually more impactful as unlocking a new way of organizing society. Finish this story:
* Communism is primitive socialism plus electricity
* The radio was an essential prerequisite for fascism
* AI will unlock ????
We read “AI as Governance” by Henry Farrell in order to understand whether and how political scientists are thinking about this question.
* Concepts or other books discussed:
* E. Glen Weyl, coauthor of Radical Markets: Uprooting Capitalism and Democracy for a Just Society, and key figure in the Plurality Institute was brought up by Seth as an example of an economist-political science crossover figure who is thinking about using technology to radically reform markets and governance.
* Cybernetics: This is a “science” that studies human-technological systems from an engineering perspective. Historically, it has been implicated in some fantastic social mistakes, such as China’s one-child policy.
* Arrow’s Impossibility Theorem: The economic result that society may not have rational preferences — if true, “satisfying social preferences” may not be a possible goal to maximize
* GovAI - Centre for the Governance of AI
* Papers on how much people/communication is already being distorted by AI:
* Previous episode mentioned in the context of AI for social control:
* Simulacra and Simulation (Baudrillard): Baudrillard (to the extent that any particular view can be attributed to someone so anti-reality) believed that society lives in “Simulacra”. That is, artificially, technologically or socially constructed realities that may have some pretense of connection to ultimate reality (i.e. a simulation) but are in fact completely untethered fantasy worlds at the whim of techno-capitalist power. A Keynesian economic model might be a simulation, whereas Dwarf Fortress is a simulacra (a simulation of something that never existed). Whenever Justified Posteriors hears “governance as simulation”, it thinks: simulation or simulacra?
Episode Timestamps
[00:00:00] Introductions and the hosts’ backgrounds in political science.
[00:04:45] Introduction of the core essay: Henry Farrell’s “AI as Governance.”
[00:05:30] Stating our Priors on AI as Governance
[00:15:30] Defining Governance (Information processing and social coordination).
[00:19:45] Governance as “Lossy Simulations” (Markets, Democracy, Bureaucracy).
[00:25:30] AI as a tool for Democratic Consensus and Preference Extraction.
[00:28:45] The debate on Algorithmic Bias and cultural bias in LLMs.
[00:33:00] AI as a Cultural Technology and the political battles over information.
[00:39:45] Low-cost signaling and the degradation of communication (AI-generated resumes).
[00:43:00] Speculation on automated Cultural Battles (AI vs. AI).
[00:51:30] Justifying Posteriors: Updating beliefs on the need for a new political science.
