Luck of the Draw III: Using Al to Extract Data About Decision-Making in Federal Court Stays of Removal
Bibliographic record
Abstract
This article examines decision-making in Federal Court of Canada immigration law applications for stays of removal, focusing on how the rates at which stays are granted depend on which justice decides the case. The article deploys a form of computational natural language processing, using a large-language model machine learning process (GPT-3) to extract data from online Federal Court dockets. The article reviews patterns in outcomes in thousands of stay of removal applications identified through this process and reveals a wide range in stay grant rates across many justices. The article argues that the Federal Court should take measures to encourage more consistency in stay decision-making and cautions against relying heavily on stays of removal to ensure that deportation complies with constitutional procedural justice protections. The article is also a demonstration of how machine learning can be used to pursue empirical legal research projects that would have been cost prohibitive or technically challenging only a few years ago-and shows how technology that is increasingly used to enhance the power of the state at the expense of marginalized migrants can instead be used to scrutinize legal decision-making in the immigration law field, hopefully in ways that enhance the rights of migrants. The article also contributes to the broader field of computational legal research in Canada by making available to other non-commercial researchers the code used for the project, as well as a dataset of several thousand Federal Court dockets that can be used for future research.
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 | 0.002 | 0.024 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.004 | 0.006 |
| Science and technology studies | 0.003 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.005 | 0.003 |
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".