Part 11/13:
Shapiro concedes that if, against all odds, a superintelligent AI emerges and becomes uncontrollable, two permanent strategies might help:
1. Decentralized Deployment
Instead of building one monolithic AI system, AI could be deployed in a distributed, networked manner.
Blockchain and DAO technologies could serve as mechanisms for consensus—AI systems would require human approval or collective agreement before taking action.
This "hive-mind" approach creates multiple points of oversight, reducing risk of rogue behavior.
2. Proper Alignment Through Core Values
- The core idea is to embed values directly into AI systems so that they inherently align with human and planetary interests.