Delegation, relinquishment, and responsibility: The prospect of expert robots
Bibliographic record
Abstract
Some day we may come to rely on robotic prediction machines in place of human experts. Google’s search engine, IBM’s Watson, and the Google Driverless Car (GDC) project each give an idea of what that world will look like. Yet actually letting go of the wheel may be a tough sell to humans. Will we really delegate human tasks to expert machine systems and what will be the outcomes of those choices? This chapter suggests that in the near future we will have to make difficult decisions about whether to relinquish some control to robots. The normative pull of “evidence-based practice” and the development of Watson-like robots will leave us few reasons to remain in control of expert decisions where robots excel. Thus we will have to choose between either accepting the relative fallibility of human experts and remaining in total control or deciding to relinquish some control to robots for the greater good. If we do relinquish some control to robots, there are important questions about the justification to do so with highly specialized expert tasks and how that would that bear on the determination of responsibility, particularly in cases of disagreement. On the other hand, if we choose to remain in control and advocate the status quo, we may deliver less than optimal outcomes relative to what “co-robotics” might achieve. Cases of disagreement between human and robot experts, generally favor delegation to robots, but also provide time for human experts to understand and make decisions on the underlying rationale for the disagreement. Watson and the GDC are able to achieve high degrees of something like “expertise” by acting on sets of rules that underdetermine their success. By describing Watson-like robots as “experts,” rather than merely “tools,” we realize a philosophical gain that accounts for both a robot’s unique abilities and social meaning.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.037 | 0.046 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.012 | 0.068 |
| Scholarly communication | 0.014 | 0.025 |
| Open science | 0.002 | 0.013 |
| Research integrity | 0.014 | 0.011 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".