Home / Library / Superintelligence: Paths, Dangers, Strategies

Superintelligence: Paths, Dangers, Strategies

🟡 Seed Entry — This entry has been imported from the Phase 1 inventory but has not yet been manually reviewed. Please cross-reference with the original source before citing.
Quantum & AI AI alignmentOrthogonality thesis

By Nick Bostrom

Summary

Intelligence is orthogonal to goals — a superintelligent AI can pursue any goal, not necessarily human-aligned ones. Consciousness is a separate question from intelligence. The "control problem" of aligning AI with human values is the central challenge of our century.

Why It Matters

The orthogonality thesis separates the question of AI consciousness from the question of AI alignment. A non-conscious AI can still be dangerous.

Notes

This entry is a seed import from the Phase 1 Knowledge Inventory. It has not yet been manually reviewed.