The problem
An institution can use a score, classification, prediction, or recommendation to structure a person’s opportunities. Explaining the outcome may still leave that person unable to identify a mistaken record, question a proxy, assess the evidence, or bring relevant circumstances into view.
The Necessary Informational Conditions (NIC) program investigates what an affected person—or a representative acting with them—needs to recognize and investigate a material objection. It also asks when the institution must provide a route through which that objection can matter.
Three connected inquiries
From Notice to Contestable Grounds
Conceptual article · OutlineWhat informational tasks must remain possible where meaningful contestability is required or promised?
B / 02Private Power without Epistemic Preemption
Normative theory · OutlineWhen does a private institution owe informed challenge and consequential reconsideration?
E / 03Keeping Independent Evidence in Play
Empirical program · ProposedTesting evidence integration and defect-matched information in synthetic credit cases.
From determination to consequential review
The common starting point is an institutionally operative determination: an output an institution adopts or materially relies on so that it shapes treatment, opportunity, visibility, priority, burden, or presumptive status.
A material objection identifies something that, if substantiated, could warrant correction, further inquiry, changed evidential weight, suspension, reconsideration, or a different account of the determination.
The organizing question
What task would a person need to perform to investigate this objection—and what would allow that task to be performed faithfully?
Paper A develops that functional question. Paper B considers the normative duty to provide answerability. The proposed studies test selected claims about information and evidence use. Each has a different burden of proof.
Six candidate informational functions
- 01
Determination and consequence
What did the institution conclude, and what practical effect followed?
- 02
Sociotechnical roles and evidence control
Who generated, supplied, interpreted, adopted, or modified the determination—and who controls its grounds?
- 03
Data, source, and quality
Are the records accurate, current, properly linked, and adequate for this use?
- 04
Governing-standard–operation fit
Does the implemented target, proxy, rule, or objective match the authorized standard?
- 05
Evidential warrant, uncertainty, and threshold
Does the evidence support the inference and the chosen level of reliance?
- 06
Contextual applicability
Do the system, evidence, and assumptions apply to this person and these circumstances?
The register is a working partition. Necessity attaches most securely to the task or a faithful equivalent, under specified conditions. The research allows functions to be merged, split, revised, or limited to particular domains.
Scope and research status
Informational adequacy does not by itself establish fairness, legality, accuracy, or legitimacy. Nor does identifying a useful information function establish a universal duty to disclose it. Institutional answerability is external to the register; it is not a seventh informational condition.
The papers are in development. The empirical work is proposed, and this site reports no completed study results. The current study plan focuses on evidence integration and two informational functions in synthetic consumer-credit cases.
Read the proposed research plan →
Selected intellectual context
The program builds on established work on contestability and reviewability. These are entry points to the surrounding literature.
- Margot E. Kaminski & Jennifer M. Urban (2021). The Right to Contest AI.
- Henrietta Lyons, Eduardo Velloso & Tim Miller (2021). Conceptualising Contestability: Perspectives on Contesting Algorithmic Decisions.
- Kars Alfrink, Ianus Keller, Gerd Kortuem & Neelke Doorn (2023). Contestable AI by Design: Towards a Framework.
- Jennifer Cobbe, Michelle Seng Ah Lee & Jatinder Singh (2021). Reviewable Automated Decision-Making: A Framework for Accountable Algorithmic Systems.