Experts Struggle to Predict AI Trajectories, But We Can Anticipate Risks from Misaligned Business Incentives
-
Leading experts often make poor predictions about the trajectory of innovations like AI. It's hard to foresee risks posed by new technologies.
-
However, we can anticipate risks stemming from companies' profit incentives misaligning with society's interests.
-
Platforms like Amazon and Google initially brought value, but later changed algorithms to maximize rents/profits over user benefit.
-
Next-gen AI like ChatGPT risks similar issues if market structure incentivizes concentration and closed ecosystems.
-
We should focus less on predicting AI risks, more on aligning economic incentives for accountability, openness and equitable value distribution.