Tech · Alignment Forum
Alignment Forum argue that the fundamental concept underlying corrigibility is respect for authority
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
◌ Single Source
Subject to this restraint, the agent then pursues the second component of its goal: acting on the principal’s behalf as the principal would judge best.
Key facts
- For example, the sentence "The temperature sensor read 80°F at 8:00 AM today" can be settled by observation at 8:00 AM as either certainly and permanently true or certainly and permanently false
- However, the most useful empirical claims are inductive generalizations (such as "the temperature is always above 70°F at 8:00 AM every day") which can never be conclusively proven via observation
- Normative feedback is permanently revisable and never fully reduces normative uncertainty to 0
- Many thanks to Abram Demski for materially and intellectually supporting this research
Summary
Many thanks to Abram Demski for materially and intellectually supporting this research. Despite the central place of corrigibility to theoretical AI alignment, it seems like no one can agree about what is. The reporter argue that the fundamental concept underlying corrigibility is respect for authority, using the analogy of a healthcare proxy's relationship to a patient as an intuitive guide. This approach stresses the deeply normative character of corrigibility, while remaining sufficiently concrete so as to be specifiable formally. Suppose you are designing a powerful AI and you need to choose which goals the AI should pursue, at whatever level of specification you can provide.