Explainability of Artificial Intelligence Models: Technical Foundations and Legal Principles
Bibliographic record
Abstract
Abstract The now prevalent use of Artificial Intelligence (AI) and specifically machine learning driven models to automate the making of decisions raises novel legal issues. One issue of particular importance arises when the rationale for the automated decision is not readily determinable or traceable by virtue of the complexity of the model used: How can such a decision be legally assessed and substantiated? How can any potential legal liability for a “wrong” decision be properly determined? These questions are being explored by organizations and governments around the world. A key informant to any analysis in these cases is the extent to which the model in question is “explainable”. This paper seeks to provide (1) an introductory overview of the technical components of machine learning models in a manner consumable by someone without a computer science or mathematics background, (2) a summary of the Canadian and Vietnamese response to the explainability challenge so far, (3) an analysis of what an ”explanation” is in the scientific and legal domains, and (4) a preliminary legal framework for analyzing the sufficiency of explanation of a particular model and its prediction(s).
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 | 0.018 | 0.039 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.002 | 0.020 |
| Scholarly communication | 0.007 | 0.011 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.004 | 0.006 |
| Insufficient payload (model declined to judge) | 0.004 | 0.000 |
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".