MAIA logo

Week 7: AI Governance & Liability

Tort Law, Compute Governance, Export Controls, and the Regulator's Toolbox

Overview

The previous weeks examined technical approaches to making AI systems safe. This session shifts to the institutional level, asking: how can governments and legal systems shape AI development and deployment? We begin with liability, exploring how existing tort law applies — and struggles to apply — to AI-caused harms, and whether the emerging patchwork of state-level legislation can provide adequate accountability. From there, we turn to compute governance as a uniquely promising policy lever, made feasible by the extreme concentration of the AI chip supply chain, and examine how the US has used export controls to maintain its lead over China in AI-relevant hardware. We close with a survey of concrete regulatory tools available across the AI lifecycle.

Learning Objectives

By the end of Week 7, fellows should be able to:

  • Explain why AI liability is difficult under existing tort law and evaluate the tradeoffs of different legal approaches (negligence, strict liability, and state-level statutory frameworks)
  • Describe how compute governance works as a policy lever — including its use for visibility, allocation, and enforcement — and articulate both its strengths and its limitations
  • Assess the effectiveness of US semiconductor export controls on China's AI ecosystem and explain the broader strategic dynamics they reflect
Reading packet

Core Readings

Recommended Readings

Further Readings

Back

In partnership with AISST

maia-exec@mit.edu
Accessibility