Q264 - Interoperability_definition_example_and_why_it_matters_in_RA
Q264 — Interoperability — definition, example, and why it matters in RAIDT
← RAIDT · Star S9 - Policy, Standards and Assurance · primary item: S9.05 · Interoperability
H. Policy, Empirical & Adoption | Ordered by mind-map priority: inner circles first, then operational detail.
Appears in sources
workshop_dense_100#slide 87
Answer
In RAIDT, interoperability means that the same governance evidence can be reused across multiple frameworks and organisational settings, while each regime keeps its own legal meaning and decision rules. The policy pathways paper defines it as the ability to compare and reuse evidence across different laws, standards, and organisations. That is why RAIDT is framed as an evidence translation layer rather than as a substitute for regulation. Its practical mechanism is simple but important: capture a run-level evidence pack for one configured use, score it through the five pillars (Responsibility, Auditability, Interpretability, Dependability, Traceability), and retain the resulting score profile so that different reviewers start from the same evidential base.
This matters because generative AI governance often fails at the last mile. A policy may say that oversight exists, a supplier may say that logging exists, and a standards programme may say that continual improvement exists, yet a reviewer may still be unable to reconstruct one contested run. Interoperability addresses that problem by making the evidence object portable. It matters in RAIDT not only for regulators, but also for procurement teams, internal auditors, domain managers, and complaint handlers. If they can all inspect the same run-level record, governance becomes more comparable, more challengeable, and less dependent on narrative assurances. In short, interoperability matters because it lets evidence travel, even when institutions, sectors, and legal regimes do not converge fully.
Practical example
Imagine a public-service eligibility system that uses generative AI to draft case explanations for staff. A claimant later challenges one explanation. Without interoperability, the authority may have a policy manual, a supplier statement, and scattered system logs, but no shared object that different reviewers can inspect coherently.
With RAIDT, the authority retains one run-level evidence pack: the exact prompt template, policy clause version, model deployment identifier, any retrieved guidance snapshot, the generated explanation, the output hash, and the reviewer sign-off. Procurement can use the same structure in supplier contracts, audit can inspect it during sampling, and policy staff can map it to transparency and oversight expectations. The example shows the core point: interoperability is not abstract harmonisation; it is the ability of one evidence pack to support several forms of scrutiny.
Sources in RAIDT papers
10-RAIDT_Policy_Pathways_M_V50