Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 60 of the 60 references it cites were found in a registry or an index; 33 of the 53 sources its claims cite were checked for support.

Reviewable Automated Decision-Making: A Framework for Accountable Algorithmic Systems

Open access
Cobbe, Jennifer, Lee, Michelle Seng Ah, Singh, Jatinder · 2021 · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency

Abstract

This paper introduces reviewability as a framework for improving the accountability of automated and algorithmic decisionmaking (ADM) involving machine learning. We draw on an understanding of ADM as a socio-technical process involving both human and technical elements, beginning before a decision is made and extending beyond the decision itself. While explanations and other model-centric mechanisms may assist some accountability concerns, they often provide insufficient information of these broader ADM processes for regulatory oversight and assessments of legal compliance. Reviewability involves breaking down the ADM process into technical and organisational elements to provide a systematic framework for determining the contextually appropriate record-keeping mechanisms to facilitate meaningful review - both of individual decisions and of the process as a whole. We argue that a reviewability framework, drawing on administrative law's approach to reviewing human decision-making, offers a practical way forward towards more a more holistic and legally-relevant form of accountability for ADM.

Overview

CoChat Summary

TL;DR

This paper introduces “reviewability” as a framework for improving accountability of automated and algorithmic decision-making (ADM) involving machine learning by breaking ADM into technical and organisational stages and specifying contextually appropriate record-keeping and logging. It argues that model-centric explanations are insufficient for regulatory oversight and legal compliance, and that a holistic, process-based approach drawing on English administrative law can better support meaningful accountability.

Key findings

  • Model-focused transparency and explanation mechanisms are insufficient because ADM is a socio-technical process with human and organisational elements; problems often originate outside the model, in purposes, assumptions, design, and use.
  • Reviewability requires contextually appropriate information that is (i) relevant to the accountability relationships involved, (ii) accurate, complete, and representative, (iii) proportionate to the required transparency, and (iv) comprehensible by the relevant forums.
  • The paper proposes a framework with four stages: Commissioning (procurement, problem definition, impact assessment), Model building (data collection, pre-processing, model training, model testing), Decision-making (deployment, use, consequences), and Investigation (audit, disclosure).
  • It draws on accountability as involving an actor, a forum, an account, and consequences, and cites five kinds of accountability relationship: political, legal, administrative, professional, and social. It also identifies three generic actor roles: managers, developers, and users.

Methods

  • Conceptual and legal-theoretical framework development, drawing on English administrative law, accountability literature, and prior work applying administrative law to ADM.
  • No empirical sample or quantitative analysis is reported; the paper synthesises related research and proposes a systematic framework for reviewable ADM.
  • Main analysis breaks the ADM process into stages and steps, then identifies record-keeping and logging mechanisms that could provide contextually appropriate information for review of individual decisions and the process as a whole.

Limitations

  • The paper states that reviewability cannot on its own address all potential problems with ADM; wider examination of political economy and socio-economic context, as well as legal protections for individuals and groups, is also needed.
  • It states that transparency has limits: reviewable systems may not capture broader business processes, wider context, interactions with other socio-technical processes, business models, or questions of power; transparency also has temporal limitations and is open to subjective interpretation and contestation.
  • It states that implementation research is needed on what record-keeping and logging are appropriate at each step, what information to retain, and how best to present it to forums; recording everything can be impractical and can create a “transparency paradox.”
  • Evident from the text: the framework is conceptual and normative, with no empirical evaluation or deployment study reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1145/3442188.3445921

Cites

60 of 60 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

23 Valid · 10 Not valid · 12 Not checked · 78 Own results and interpretations

24 more claims · 8 with their cited sources checked

Create a free account to see all 123 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 87
  • It may state two separate claims as one. 24

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 52 references

Research works printed but not linked: 14

  1. Philip Alston. 2019. Report of the Special Rapporteur on extreme poverty and human rights. United Nations General Assembly A/74/493 (2019). https://undocs.org/A/74/493
  2. Mark Bovens. 2006. Analysing and Assessing Public Accountability. A Conceptual Framework. EUROGOV, European Governance Papers No. C-06-01 (2006).
  3. Cary Coglianese and Daniel Lehr. 2017. Regulating by Robot: Administrative Decision Making in the Machine-Learning Era. Georgetown Law Journal 105 (2017).
  4. European Commission. 2020. White Paper On Artificial Intelligence - A European Approach to Excellence and Trust. COM(2020) 65 (2020).
  5. 10.1145/2347736 · doi:10.1145/2347736
  6. Avi Feller, Emma Pierson, Sam Corbett-Davies, and Sharad Goel. 2016. A computer program used for bail and sentencing decisions was labeled biased against blacks. It's actually not that clear. The Washington Post (2016).
  7. Agata Foryciarz, Daniel Leufer, and Katarzyna Szymielewicz. 2020. Black-Boxed Politics: Opacity is a Choice in AI Systems. Panoptykon Foundation (17 January 2020). https://en.panoptykon.org/articles/black-boxed-politics-opacity-choice-ai-systems
  8. 10.1093/idpl · doi:10.1093/idpl
  9. David Lehr and Paul Ohm. 2017. Playing with the Data: What Legal Scholars Should Learn About Machine Learning. UC Davis Law Review 51 (2017).
  10. House of Common Science and Technology Committee. 2018. Algorithms in Decision-Making. Fourth Report of Session 2017-19 HC 351 (2018).
  11. Douglas Pyper. 2020. Research Briefing: The Public Sector Equality Duty and Equality Impact Assessments. House of Commons Library (2020). https://commonslibrary.parliament.uk/research-briefings/sn06591/
  12. Nick Seaver. 2013. Knowing Algorithms. paper presented at Media in Transition 8 (2013).
  13. Cynthia Stohl, Michael Stohl, and Paul M Leonardi. 2016. Managing Opacity: Information Visibility and the Paradox of Transparency in the Digital Age. International Journal of Communication 10 (2016).
  14. European Union. 2016. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation). Official Journal of the European Union L119 (4 May 2016), 1--88.

Cited by

See all 150

Metadata

Authors
Cobbe, J., Lee, M. S. A., & Singh, J.
Affiliations
University of Cambridge (GB)
Venue
Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency
Publisher
ACM
Published
2021-03
Pages
598-609
OpenAlex
W3134395196
Cited by
150
References
52
Citations by year
2026: 63 · 2025: 29 · 2024: 15 · 2023: 13 · 2022: 18 · 2021: 11
FWCI
12.86
Type
conference paper
Language
en
Open access
Locations
ACM Conference on Fairness, Accountability, and Transparency (FAccT) (publishedVersion, cc-by)
Funders
Engineering and Physical Sciences Research Council (EP/P024394/1, EP/R033501/1)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Law › Digitalization, Law, and Regulation
  • Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
Keywords
reviewability, automated decision-making, algorithmic decision-making, accountability, machine learning, administrative law, regulatory oversight, legal compliance, record keeping, socio-technical processes, administrative review

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-29, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-29 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Reviewable Automated Decision-Making: A Framework for Accountable Algorithmic Systems. https://app.cochat.ai/papers/10.1145/3442188.3445921
The paper
Cobbe, J., Lee, M. S. A., & Singh, J. (2021). Reviewable Automated Decision-Making: A Framework for Accountable Algorithmic Systems. Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency. https://doi.org/10.1145/3442188.3445921

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.