Mitigating Bias Due to Race and Gender in Machine Learning Predictions of Traffic Stop Outcomes
Bibliographic record
Abstract
Traffic stops represent a crucial point of interaction between citizens and law enforcement, with potential implications for bias and discrimination. This study performs a rigorously validated comparative machine learning model analysis, creating artificial intelligence (AI) technologies to predict the results of traffic stops using a dataset sourced from the Montgomery County Maryland Data Centre, focusing on variables such as driver demographics, violation types, and stop outcomes. We repeated our rigorous validation of AI for the creation of models that predict outcomes with and without race and with and without gender informing the model. Feature selection employed regularly selects for gender and race as a predictor variable. We also observed correlations between model performance and both race and gender. While these findings imply the existence of discrimination based on race and gender, our large-scale analysis (>600,000 samples) demonstrates the ability to produce top performing models that are gender and race agnostic, implying the potential to create technology that can help mitigate bias in traffic stops. The findings encourage the need for unbiased data and robust algorithms to address biases in law enforcement practices and enhance public trust in AI technologies deployed in this domain.
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 | 0.028 | 0.077 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 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".