Qualification: Conceptualising Contestability: Perspectives on Contesting Algorithmic Decisions

Record: kaal:position:2026-08-26-026 · 2026-08-26

Explanation begins a challenge. It does not complete one. Lyons, Velloso, and Miller examine contestability in algorithmic decision making through a thematic analysis of 65 submissions to Australia's AI Ethics Framework. They distinguish information from institutional process. An explanation supplies grounds to question a decision. Contestability also requires notice, a clear path into review, a reviewer with authority to act, and a new decision or outcome when review is complete. The paper identifies a basic failure condition: users may receive a nominal right to appeal and still receive no response or resolution. The remedial element is concrete. The reviewed submissions describe redress as a new decision, compensation for harm, or reversal of a decision's effects. They connect the availability of those remedies to enforceability. Internal review has value because the responsible institution may be positioned to reverse its decision or provide redress. External review remains necessary where independence or escalation controls. This evidence qualifies the institutional requirement for sovereign local agent runtimes. A runtime may record an explanation and expose a complaint interface without creating a functioning institution. The challenge path must identify what can be contested, who can decide the challenge, what remedial powers are available, and how a resulting decision becomes effective. Explanation supports that process. It cannot substitute for remedial capacity. The study does not test a deployed agent runtime. Its evidence consists of stakeholder submissions and the authors' qualitative synthesis, not observed remedy rates. It does not establish which reviewer, remedy, timing rule, or enforcement mechanism is sufficient for every context. It also warns that contestation places burdens on affected persons and cannot replace lawful design, oversight, audit, or prevention. The narrower implication remains. A challenge mechanism without authority to change consequences is an information channel, not an institutional remedy.

Affirmed commentary position. This record extends a source-bound scholarly claim but is not a verbatim paper claim.
Holds when
Current debate

Conceptualising Contestability: Perspectives on Contesting Algorithmic Decisions

Scholarly basis

kaal:claim:7314479-026
Wulf A. Kaal, Institutional Requirements for Sovereign Local Agent Runtimes (2026). SSRN: https://ssrn.com/abstract=7314479
Source PDF sha256: debace24a155ae924a155b1fafe98856d98cf83689feff2f87a32f1c06171ce6

Evidence and mapping

Evidence: peer-reviewed ACM journal article with complete author-posted full text and empirical thematic analysis
Review tier: independent substantive scholarly-growth qualification
Mapping confidence: 0.98
Mapping ambiguous: false

Topics

institutional-designgovernance-designai-and-agentsaccountabilitycontestabilityreviewremedyalgorithmic-decision-making

Provenance

Affirmed in kaal-review:2026-08-26:scholarly-growth-7314479-026-reviewed-v1 on 2026-08-26. Review record.

Verify

Canonical markdown sha256: 2d2a12d3f4ff4b3277cea48e7cfb80e56bf2c3ec009f01077037344f88659af5
curl -s https://wulfkaal.github.io/positions/2026-08-26-026.md | sha256sum