Contribution to the handbook
Priority · Assessed 2026-10-27
The question distinguishes human-selected searches, algorithm-discovered patterns, human review or action, and full automation. Those choices materially affect authority, accountability, and the possibility of challenging recommendations, so they can inform system specification and safeguards.
Does this challenge the goal?
In-scope gap
The question specifies possible divisions of work between people and algorithms within the accepted governance-design goal; it does not contest that goal.
The source describes possible arrangements and an illustrative tax-enforcement example, not a universal taxonomy or evidence that any arrangement is effective.
Why this category?
- material system specification across levels of human and algorithmic control
- authority and accountability implications
- implementation and reviewability limits
The cited material is largely an account of possible roles and illustrative cases; it does not demonstrate how often each arrangement occurs or how well human review works in practice.
This is an AI assessment of contribution to the goal, separate from whether the question has an answer.
- What roles can people and algorithms have in data mining?
- Systems can learn rules from data
- How algocracy differs
- People need to examine the factors
- Algorithms organize decision work
- Review and possible override
- Human involvement may not be enough
- Algorithms may help find tax audit leads
- Better performance, less understanding
- Some systems are hard to explain
- Human control in tax enforcement
- Some systems resist review
- People can set or leave the search
Judging situations and resolving disputes →
Authority, accountability and resilience →
Tests, stop conditions and unresolved work →
Read the exploration goalExplain how people and algorithms can divide the work of selecting or finding patterns and of reviewing, acting on, or automating the resulting recommendations.
What we can say so far
Danaher describes several ways the work can be divided. People can specify which patterns a data-mining system should search for, or allow the algorithm to find patterns itself—including by learning rules from data. In the tax-enforcement example, an official may request algorithmic selection of audit targets, while an algorithm may also search data and recommend audits. Afterward, people may review recommendations, accept or reject them, and act on them; alternatively, the process or its outputs may be left to machines. These are possible arrangements, and systems may combine them rather than fit one category. The account describes roles, not how common or effective each arrangement is.
What remains open
- The material does not establish how common each arrangement is or provide a general test for whether human review is practically meaningful.
Conditions and limits
- The account presents possible roles and examples, not evidence about how frequently particular arrangements occur or their effectiveness.
- People’s formal involvement does not necessarily mean they can meaningfully understand or challenge a system.
- The evidence is based on an accessible mirror of the author manuscript, not the publisher typeset version.
Objections to consider
- A person may be formally included in review yet unable to challenge a system they cannot understand. This qualifies the distinction between nominal review and meaningful review; the supplied evidence does not resolve the limitation.
- Opacity or dependence on interconnected systems may prevent review. The evidence identifies these barriers but does not establish when they can be overcome.
- A human-machine arrangement does not necessarily mean that the person meaningfully understands or guides the decision. The passage raises this concern without resolving its application beyond the example.
- The account suggests that reviewing recommendations may require examining the factors behind them, but does not specify what access is sufficient for effective scrutiny.
- The author allows that algorithmic systems could support participation but says this may be difficult in practice. The passage does not establish how often such designs work or whether the barriers can be addressed.
Evidence and related ideas
Supporting evidence, qualifications and context play different roles in this answer.