The “Info Finance” Alternative 🔮

[etfs]

Instead of naive AI governance, Vitalik Buterin proposes a different model: info finance.

How it works:

Open market for multiple AI models to contribute insights

Spot-checks and human juries validate results

Prediction markets reward participants who surface accurate information

Benefits:

Diversity of models, not a single centralized AI

Incentives to detect and correct manipulation quickly

Reduces the risk of one compromised model controlling governance

Buterin calls this an “institution design” approach: open, competitive, and incentive-aligned, rather than relying on a single AI overlord.

[Nahid Hossen]

Vitalik’s “info finance” sounds like a reality check. Don’t trust one AI oracle—let models compete and let markets decide who’s actually accurate.

[Abdul Khan]

Prediction markets + human juries could be the missing piece. Instead of blind trust in AI, we get incentives for truth and penalties for manipulation.

[Rimon Khan]

AI governance fails when it’s centralized. Info finance makes it plural—many models, open competition, aligned incentives. Much harder to corrupt.