Asymmetrical Governance: Auditing Algorithms to Preserve Due Process Rights
Bibliographic record
Abstract
We are now living in age where algorithms, and the data that feed them, govern a wide variety of decisions in our lives: not just search engines and personalized Netflix suggestions, but educational evaluations, stock market trades and political campaigns, the urban planning, and even how social services like welfare and public safety are managed. Heterogeneous lists like this have become the norm in any critical examination of algorithms, giving the impression of a ubiquitous relevance of algorithms. But algorithms can make mistakes that directly affect individuals and often contain both implicit and explicit biases. The technical complexity of algorithms, the scale at which they operate, and their proprietary nature makes them difficult to scrutinize, creating challenges to fully comprehend of how they exercise their power and influence over society. When used to make legal decisions, questions must be asked as to how automated decision-making systems affect the right to due process afforded to citizens.\nThe goal of this research project is to augment this discussion by focusing on algorithmic transparency, due process rights, and what can be done to help protect said rights when automated decision-making systems are used. Therefore, the research question that guides this paper is as follows: In what ways do algorithms in legal processes negatively impact an individual’s right to due process and how might the ability to audit legal algorithms help protect due process rights?\nTo answer this question this research project will present recommendations for research methods, adapted from Communication scholar, Christian Sandvig’s proposed research methods (2014) that can be utilized to audit algorithms so as to provide greater transparency to those on the receiving end of algorithmic judgements within the legal process.\nAsymmetrical Governance:\nAuditing Algorithms to Preserve Due Process Rights
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 | 0.161 | 0.469 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.005 |
| Science and technology studies | 0.005 | 0.031 |
| Scholarly communication | 0.017 | 0.037 |
| Open science | 0.004 | 0.010 |
| Research integrity | 0.004 | 0.007 |
| Insufficient payload (model declined to judge) | 0.008 | 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".