The AI industry is widely considered to be in an 'arms race', such that speed is incentivised even while incurring great risks. Model this situation in more detail, assessing under what circumstances speed is desirable for parties.
About the project
Model the details of interaction between parties building AI, to better determine whether the present situation is well considered as an 'arms race', i.e. where fast development is the selfish optimum for parties.
Very preliminary modeling described and linked to here: https://worldspiritsockpuppet.substack.com/p/lets-think-about-slowing-down-ai
Theory of change
If AI is not an arms race, then many people are forwarding a technology they believe to pose an extreme risk to society out of a mistaken belief that this will have worthwhile payoffs in expectation. Careful analysis of the situation could prevent this harm.
Your role
Mentees will run and carry out the project, with guidance from me.
Prerequisites
- Writing proficiency
- Familiarity with basic game theory
Location preference
10am-6pm PST
Application question(s)
- Provide a link to a writing sample
About the mentor
Katja Grace co-founded AI Impacts in 2014. She runs the Expert Survey on Progress in AI. She blogs at worldspiritsockpuppet.substack.com. She was considered by TIME magazine to be one of the 100 most influential people in AI in 2024.