S1.05 - IS_governance_origin

S1.05 ? IS governance origin

flowchart LR
    A[IS governance tradition
decision rights, accountability,
controls, socio-technical oversight] --> B[RAIDT - run-level evidence framework]
    H[GenAI operational pressure
need for practical governance] --> B
    B --> C[[IS governance origin
governance-first framing for GenAI runs]]
    K[Practical fields
user role, task, prompt,
model/version, approvals,
logs, exceptions] --> C
    C --> D[Evidence pack]
    C --> E[RAIDT score profile]
    C --> F[Reviewability and contestability]
    D --> G[Reviewer reconstruction]
    E --> I[Governance readiness]
    F --> J[Organisational learning and policy alignment]

← Star S1 - Origins, Background and History

Star context: Explains why RAIDT emerges not only from Responsible AI debate, but from the Information Systems governance tradition of decision rights, accountability, control, oversight and socio-technical coordination in organisational work.

Academic picture

Definition / background

IS governance origin refers to the fact that RAIDT is rooted in the Information Systems governance tradition rather than in technical optimisation alone. In that tradition, the central questions are who has decision rights, how accountability is allocated, what controls apply, how oversight is exercised, and how technical systems remain aligned with organisational purpose. RAIDT inherits those questions and applies them to generative AI use in organisational settings.

This origin matters because the governance problem created by GenAI is not only whether a model can produce an output. The deeper issue is whether a specific use of that model happened under the right authority, with the right configuration, for an appropriate purpose, with reviewable evidence, and with a defensible human accountability structure. IS governance provides the conceptual language for that broader problem.

RAIDT therefore belongs inside IS governance because it treats the run as the unit at which organisational control becomes visible. The run-level evidence pack documents the socio-technical circumstances of use, while the five-pillar score profile translates those circumstances into a structured governance view across Responsibility, Auditability, Interpretability, Dependability and Traceability. This is different from generic AI governance language that often remains at policy level or model level without showing how a particular organisational use can be reconstructed and reviewed.

Why this concept matters

Without an IS governance origin, RAIDT could easily be misunderstood as a technical benchmarking tool, a compliance checklist, or a Responsible AI statement with added metadata. The concept matters because it clarifies that RAIDT is designed to govern the use of GenAI inside organisational processes. It frames GenAI outputs as part of decision structures, review pathways, accountability chains and control environments.

This helps avoid a common confusion in AI governance: the assumption that better governance is mainly a matter of better models or more principles. Organisations also need a way to show who used what, for which task, under which conditions, with what review path, and how that use can later be challenged or improved. If this governance lineage is missing, organisations risk adopting GenAI in ways that are technically impressive but operationally opaque, weakly controlled and difficult to defend.

Key idea: IS governance origin matters because it places RAIDT inside the tradition of accountable organisational control, then makes that control operational through run-level evidence.

What this item explains

Why RAIDT should be understood as a governance framework for organisational GenAI use rather than only a technical AI assessment tool.
How decision rights, accountability and control structures attach to a specific GenAI run.
Why evidence for governance must include workflow context, human roles, approvals and exceptions, not only prompts and outputs.
How run-level documentation supports reviewability and contestability when a GenAI-assisted action is questioned.
Why the five RAIDT pillars express governance quality in an operational form rather than as abstract principles.
How IS governance origin links RAIDT to audit readiness and organisational learning over time.

Practical example / likely audience question

Audience question

Is RAIDT an IS project or a technical AI project?

Answer

The concern behind this question is that governance frameworks are often treated as either technical or managerial, as if they must belong to one side only. RAIDT is primarily an IS governance project because its central object is the governed use of a system inside organisational work. Technical evidence remains important, but it is used in support of governance judgement rather than as a substitute for it.

A practical example is a university using GenAI to draft first-pass student-support case summaries. The key issue is not only whether the model can produce coherent text. The real governance questions are who may use the tool, what data are permitted, what template and model version were used, what human review is required, where the output is stored, and how a disputed summary can later be reconstructed. RAIDT handles this better than a generic AI governance approach because it ties those questions to a specific run, a specific evidence pack and a specific score profile, rather than leaving them as high-level policy claims.

Practical example in RAIDT terms

A retail bank uses GenAI to produce first-draft summaries for SME loan applications. The run-level issue is not simply whether the model can summarise documents accurately. The issue is whether one particular run used approved inputs, an authorised prompt template, the correct model and version, a named analyst, a defined reviewer, and a clear boundary that the lending decision remains human.

The evidence needed would include the business purpose of the run, the identity and role of the user, the model and version, prompt template identifier, data classification, source documents used, approval checkpoint, reviewer sign-off, exception log and retained output. The most affected RAIDT pillars are Responsibility, because ownership and review must be explicit; Auditability, because the run must be reconstructable; Interpretability, because the rationale for using the output must be understandable; Dependability, because the workflow must operate consistently; and Traceability, because the output must be linked back to inputs and decision records.

IS governance origin improves governance readiness here because it positions the run inside a controlled organisational process. The GenAI output is not treated as a free-floating technical artefact. It is treated as an event inside a lending workflow that must be authorised, reviewable and contestable.

Detailed link to RAIDT

IS governance origin links to RAIDT in four ways.

First, it links to RAIDT's core idea by defining GenAI governance as an organisational control problem, not merely a model-performance problem.

Second, it links to the run because a run is the point at which decision rights, user roles, configuration choices, workflow context and review obligations become concrete and inspectable.

Third, it links to the evidence pack and score profile because those outputs make governance claims demonstrable. The evidence pack records the relevant socio-technical facts of use, and the score profile translates them into a structured judgement across the five pillars.

Fourth, it links to reviewability, contestability, audit readiness and organisational learning because a governance framework becomes credible only when reviewers can reconstruct what happened, question it, compare cases and improve controls over time.

IS governance origin -> Run-level evidence -> Evidence pack -> RAIDT score profile -> Governance readiness

This chain shows why the item matters. IS governance origin gives RAIDT its framing, run-level evidence gives it granularity, the evidence pack gives it documentary substance, the score profile gives it evaluative structure, and governance readiness gives it organisational purpose.

Link to the five RAIDT pillars

Responsibility

IS governance origin strongly affects Responsibility because it starts from the allocation of decision rights, ownership and answerability in organisational work. A run should always sit within a named authority structure rather than an informal pattern of tool use.