Accessible Contestability in Algorithm-Assisted Adult Social Care: Legal Limits and Governance Duties in Scotland after the Data (Use and Access) Act 2025
Year: 2026
Type: preprint
Abstract: Artificial intelligence is already present in Scottish social care. Documented uses include AI-assisted transcription and assessment drafting in local-authority social work, facial-analysis tools for pain assessment in care homes, machine-learning monitoring in supported housing, and developing AI-supported referral triage. Public evidence does not yet establish the prevalence or institutional distribution of these systems. It does establish that the legal problem is current rather than hypothetical. This article argues that the principal risk is not confined to a machine taking the final decision. AI systems increasingly shape the evidential record, the classification of need, the priority assigned, and the information presented to the human decision-maker. Meaningful human involvement must therefore include verification of AI-generated records and not merely human approval of a final outcome. The Data (Use and Access) Act 2025 provides an important statutory framework through new Articles 22A to 22D of the UK GDPR. Its safeguards remain incomplete for systems that materially influence a decision while a human officer remains formally involved. Scottish adult social care provides a demanding setting in which to address that gap. It combines devolved assessment and self-directed support duties, integrated delivery, equality law, digital exclusion, complaints architecture, and decisions with immediate consequences for autonomy and daily life. The article develops an Accessible Contestability Framework that ranks each element by legal strength. Its nine elements are accessible notice, explanation sufficient for challenge, meaningful human reconsideration, record provenance and verification, correction and contextualisation of data, multi-channel access, support and advocacy signposting, complaint and redress signposting, and audit and procurement evidence. The framework is not presented as a new freestanding legal rule. It is a disciplined account of how existing legal duties and governance standards should constrain both automated decisions and AI-mediated evidential production in Scottish adult social care.
Author: Koesters, Nils
Language: English
Subfield: Law
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7166459354
Cites:
Cited by:
Related to: