Well-Founded Fear of Algorithms or Algorithms of Well-Founded Fear? Hybrid Intelligence in Automated Asylum Seeker Interviews
Bibliographic record
Abstract
Abstract Growing numbers of asylum seekers across Europe have created heightened pressure on governments to employ technologies to assist immigration systems in meeting humanitarian standards of international law. This article analyses the potential of hybrid intelligence (HI)—a machine learning (ML) utility supervised by and supervising human intelligence—for assisting both asylum seekers and immigration officers in performing fair and just assessments, while addressing theoretical underpinnings of what hybridity entails from the perspective of stakeholders and humanitarian systems. While aspects of ML demonstrate promise in reducing bias in immigration decisions, such technology itself suffers from various inherent biases. In addition, technological mediation poses several unforeseen, unintended, and subtle threats to humanitarian missions. By analysing ML algorithms currently employed in refugee status determination pilot programs and immigration control, this article synthesizes universal complications of using assistive technology in Refugee Status Determinations, with special focus on evaluating resultant theoretical refugee identity reconfigurations. Conceptually, this article expands on the theoretical model of what has been termed ‘ID entity’ by biometrics researchers and ethnographers by analysing potential latent consequences from technological mediation in asylum cases, while addressing use cases such as German and Canadian immigration services’ pilot programs, along with automated pilot border screening projects such as Iborderctrl, among others. In addition, several hypothetical scenarios are presented to concretize and further theoretical inquiry of using HI in asylum seeker interviews, with special focus on the requisite criterion of possessing a well-founded fear of persecution.
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 | 0.025 | 0.051 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.006 | 0.021 |
| Scholarly communication | 0.007 | 0.009 |
| Open science | 0.002 | 0.008 |
| Research integrity | 0.002 | 0.004 |
| 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".