AI governing AI
An approach to AI safety where one AI system is used to monitor, evaluate, and correct the outputs of another AI system, operating on the assumption that AI can regulate AI.
As defined in Jaron Lanier Looks into AI's Future | AI IRL
Where it comes up · 1
Most important claims · 2 of 2Search all claims for “AI governing AI” →
- AI-governing-AI governance models fail because language itself is fundamentally ambiguous—words can always be twisted, as illustrated by the ancient genie problem where wishes are always subverted by semantic twisting, and this problem is mathematically irreducible.Jaron Lanier Looks into AI's Future | AI IRL
- Regulation of AI should be based on data provenance rather than AI-governing-AI approaches, because the latter creates infinite regress and relies on undefined terms like 'alignment with human values' that are too ambiguous to enforce.Jaron Lanier Looks into AI's Future | AI IRL