Contribution to the handbook
Priority · Assessed 2026-10-27
The question highlights that presumed knowledge advantages may not translate into explainable reasoning. The evidence also raises difficulty understanding complex systems and the possibility that review may depend on experts, making this material to scrutiny and accountability.
Does this challenge the goal?
In-scope gap
The question explores a practical limitation on understanding and scrutiny within the existing governance goal; it does not contest that goal.
The cited accounts describe possible sources of opacity and normative concerns, but do not establish how often they occur or what level of explanation is required.
Why this category?
- material system specification
- consequential risks and limits
- implementation dependency
These sources share roots where noted and are interpretations of source material, not independent evidence that opacity is inevitable or that a specific remedy works.
This is an AI assessment of contribution to the goal, separate from whether the question has an answer.
- Why might a group’s decision-making remain hard to explain, even when it is thought to know better?
- Knowledge may be hard to explain
- Some AI systems may exceed human understanding
- Some results are hard to explain
- Recording alone may not explain a decision
- Understanding may be the deeper problem
- People need to understand the process
- The argument and its elite
- Reasons for choices are unclear
Authority, accountability and resilience →
Making and interpreting laws →
Judging situations and resolving disputes →
Tests, stop conditions and unresolved work →
Read the exploration goalExplain possible gaps between a group’s presumed knowledge advantage and its ability to articulate and communicate its reasoning.
What we can say so far
A group’s presumed knowledge advantage does not guarantee that it can explain its reasoning. The paper describes several possible gaps: members may know more than they can put into words; they may have difficulty communicating their reasoning to the wider public; and an advantage arising from the group’s circumstances may be hard to identify or explain. People may also forget why the group was originally favored while continuing to defer to it. In algorithmic settings, a further possible obstacle is that complex or hard-to-interpret processes can make the route from information to a decision difficult to understand. These are proposed possibilities, not claims that such opacity always occurs.
What remains open
- The supplied material does not establish when these problems arise in practice or how often they occur.
- It does not quantify how much any one mechanism impairs public understanding or communication.
Conditions and limits
- The central explanation is attributed to Danaher’s argument as represented in the accessible author-manuscript text, not presented as an independently verified empirical finding.
- The mechanisms are described as possible problems, not inevitable outcomes.
- Algorithmic complexity and machine-learning interpretability evidence concerns systems and does not by itself establish the same mechanism for every knowledgeable group.
Evidence and related ideas
Supporting evidence, qualifications and context play different roles in this answer.
supports
Knowledge may be hard to explain The paper identifies several ways even a supposedly knowledgeable group may fail to make its reasoning clear: its members may know more than they can explain, may struggle to explain it to the wider public, or may be unable to describe an advantage that arose from group circumstances. People may also forget why the group was favoured yet continue to defer to it.
Directly supplies the central mechanisms: knowledge that members cannot explain, difficulty explaining to the public, circumstantial group advantages that are hard to describe, and continued deference after reasons are forgotten.
The Threat of Algocracy: Reality, Resistance and Accommodation
Read full source: The Threat of Algocracy: Reality, Resistance and Accommodation (opens in a new tab) ↗