Symbolic representation of the challenge in improving AI systems
General Definition
1What is Recalcitrance?
Recalcitrance refers to the quality of being obstinately defiant of authority or restraint, or the trait of being difficult to manage or control. It essentially describes a stubborn unwillingness to comply or cooperate.
Source: General definition
As depicted above but within AI it has its own meaning.
AI-Specific Definitions
2Recalcitrance in AI Context
Recalcitrance is defined as the inverse of the rate of increase to intelligence per time per unit of effort put into increasing that intelligence.
Source: Digifesto (4th item down)
3Practical Application
An AI system's Recalcitrance can be defined as the resistance to improvements in model accuracy in relation to the efforts put into maximizing its prediction accuracy.
Source: Sriharsha Ganjam on LinkedIn
4Core Concept
It is meant to capture how hard it is to make an intelligent system smarter, and in particular how hard it is for an intelligent system to make itself smarter.
Key Insights
- Recalcitrance measures the difficulty of improving AI intelligence
- High recalcitrance means diminishing returns on improvement efforts
- It's particularly relevant for the concept of AI self-improvement
- Understanding recalcitrance is important for AI safety research
5Further Reading
For a detailed analysis of recalcitrance and its implications for superintelligence:
Recalcitrance examined: an analysis of the potential for superintelligence explosion
