A Method to Test the Ethics of Some AI Classifiers - The Example of School Dropouts Problem
Bibliographic record
Abstract
If an AI artifact is an emulation of human behavior in relation to the performance of some activity and if the human being, in carrying out that activity, is required to respect a behavioral framework defined by laws, rules, regulations, procedures, best practices, etc., then the AI that emulates that human behavior is also required to respect the same behavioral framework. The idea of ethics tests is developed on this principle and, precisely on the basis of this principle, apragmatic methodology can be developed that can test the correspondence in the observance of the artefact to the behavioral framework within which it will necessarily be placed in its operation. The approach proposed in this work allows us to offer an extremely pragmatic solution to the search for an “ethical behavior” for AI artifacts, bypassing the difficult applicability of the complex and abstract legislation currently in force on this topic. In order to explain the applicability of this methodology to a concrete problem, this work considers theproblem of school dropout as an example and describes theconstruction of two classifiers, one based on a neural network and one on a decision tree, able to predict the phenomenon. The application of the methodology clearly shows how the explainability offered by a symbolic system, such as a decision tree, is not applicable as an element of explainability in the behavior of a neural classifier.
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 | 0.009 | 0.028 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.003 | 0.003 |
| 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".