Follow the link for the full summary: bcf.princeton.edu/events/luis...
On Thursday, April 25, Luis Videgaray joined Markus’ Academy for a conversation on “The Race to Regulate AI.” Luis Videgaray is a Senior Lecturer at MIT Sloan and the Director of MIT AI Policy for the World Project. He is also the former Foreign Minister and Finance Minister of Mexico.
A summary in four bullets
- AI is hard to regulate because of the speed of innovation and because it is a general purpose technology, so it is difficult to envision the different use (and misuse) cases.
- When designing AI laws, policymakers should: choose between a horizontal or a vertical approach, decide whether to establish a new AI regulator, and (most critically) decide whether to regulate the inputs to models or their outcomes.
- After recognizing the difficulty of developing LLMs, emerging markets should prioritize regulating their application and customization. The priority should be to demand guarantees from LLM makers and to establish clear liability rules.
- We should be optimistic about the power of AI to improve the workings of government in emerging markets, for example by improving processes like the allocation of construction permits or the distribution of welfare benefits.
Timestamps:
[0:00] Introduction and poll questions
[10:59] The race to regulate AI
[28:19] Designing AI Laws
[37:28] Why is it so hard to regulate AI?
[47:55] Q&A