A5 — Board / Council Reporting One-Pager
One page for reporting AI readiness in research training to a governing body: five dimension levels, not one misleading number.
Browse the full online instrument catalogue to read or download every resource in HTML, Word, PDF, and spreadsheet formats.
About this instrument
| Purpose | Turn the A4 reconciled profile into one reportable page: five dimension levels, each binding axis, the constraint that follows, three asks, and the regulatory exposure. |
| Who completes it | The named Dimension 5 owner, countersigned by the DVC-R, PVC-R, or VP Research who presents it. |
| Time required | 30 minutes, once A4 is complete. This page reports the profile, but it does not produce one. |
| Report sections | §1.3, §2.1–§2.5, §4.1–§4.3, §5.1, and Appendices A and F. |
| Cells touched | Summarizes all 20 cells. It scores none of them, because documentary scoring begins in A1 and the evidence is reconciled in A4. |
| Cadence | Every reporting cycle. The Since last report column makes it repeatable across cycles. |
Table 1. Summary of the purpose and structure of the board reporting one-pager instrument.
How to use this page
Only the block between the horizontal rules is presented to the governing body. Everything here is preparation.
- Filling the profile table. For Level, record Absent, Nascent, Established, Leading, or Unresolved/not demonstrated. Use the unresolved status only when an unanswered nascent indicator prevents an axis cell from receiving a maturity level. For Binding axis, record policy, people, systems, or process. If the level is unresolved, list the unresolved axes instead and leave the binding dimension unassigned. For Since last report, record "Unchanged", or a change such as "Nascent → Established". For What holds it there, write one clear clause naming the missing artifact or capability.
- Levels and asks come from A4. Use the reconciled profile rather than copying the initial A1 documentary level. Each ask should be a priority work package toward a target rung in the binding cell. Completing that work package triggers a rescore of all applicable A1 indicators at the target rung and below. It does not move the cell or dimension automatically. Replace the three default asks with the ones your institutional profile requires.
- Write block 2 fresh each cycle. You should follow this general structure: "Dimension 3 sits at nascent, with the binding axis being systems. We ask researchers to verify AI-generated citations against the primary source, but with no procurement gate, the tools they use do not emit citations that resolve. Until that gate exists, our Dimension 1 and 2 rules cannot operate above nascent in practice, whatever the written policy says."
- Levels advance progressively, and a cell moves only when supported by a dated artifact. Advancing along a dimension requires demonstrable achievement across all cells at that level, so an institution cannot reach leading while failing to meet the requirements for established (Appendix A). If an unresolved nascent rung prevents a dimension from being calculated, report it as Unresolved/not demonstrated even when overall documentary coverage is 80 percent or higher. If A1 documentary coverage is below 80 percent, the coverage verdict must be Provisional and all unanswered cells must be named explicitly. Coverage and maturity determination are separate.
- To chart the profile, plot determined levels only (absent 0, nascent 1, established 2, leading 3), label each with its binding axis, and mark any Unresolved/not demonstrated dimension separately rather than assigning it a numeric value. Do not plot a total. Never publish a composite score, because the diagnostic value lies in the pattern across dimensions rather than a single uninformative number (§4.3).
- Do not attempt to re-score at the board table. If the governing body finds itself debating an axis rating during the meeting, that is usually a sign that the underlying A4 evidence and resolution note need further work.
- Class A to D policy scope (§1.3) is a separate descriptor that is never derived from these readiness levels. If the board or council asks "what class are we?", answer directly from the policy-scope taxonomy in Responsible AI in Academic Research: A Competency Framework for Research Training.
- Other jurisdictions should substitute their own equivalent regulatory pressure points in block 4. The underlying axes and the overall reporting format remain the same regardless of location (Appendix F). An institution with both an Australian footprint and European Union market activity should report both lines.
- For the first report: write "First report, baseline established [DATE]" and leave the Since last report column empty.
- Reaching leading on Dimension 5 requires published scoring with an explicit method and external benchmarking or peer audit conducted at a stated cadence (Appendix A, row 5). Publishing internal scores alone is not enough. Section 4.2 of the report describes this as a forward-looking standard that early adopters can choose to work toward. This is worth explaining to the board or council, but it is not a reason to inflate the institution's current level.
AI readiness in research training: [INSTITUTION NAME], report to [COMMITTEE], [DATE]
1. Readiness profile
| Dimension | Level | Binding axis | Since last report | What holds it there |
|---|---|---|---|---|
| D1: Human-in-the-loop discipline | [ ] | [ ] | [ ] | [ ] |
| D2: Responsible use in practice | [ ] | [ ] | [ ] | [ ] |
| D3: Tooling that promotes responsible use | [ ] | [ ] | [ ] | [ ] |
| D4: AI-literate humans | [ ] | [ ] | [ ] | [ ] |
| D5: Institutional benchmarking grid | [ ] | [ ] | [ ] | [ ] |
Table 2. Readiness levels, binding axes, recent progress, and key constraints across institutional dimensions.
Cells improved since [DATE OF LAST REPORT]: [ ] of 20 · unchanged [ ] · regressed [ ] · coverage [ ]%.
Evidence base: A1 documentary coverage [ ]% · A2 response [ ]% (N = [ ]) · A3 response [ ]% (N = [ ]) · D5 assurance [assured / unassured].
Reading the table. Each dimension takes the lowest of its four axes (policy, people, systems, process) because reaching established requires meeting standards across all four axes at a regular, reviewable cadence (§4.1). That lowest axis is the binding one that holds the dimension back. We do not report an overall total score.
2. Binding dimension and what it constrains
Binding dimension: [D_n: name]. Binding axis: [axis].
[Two sentences explaining what capability is absent, and which other dimension it disables in practice.]
An institution's weakest dimension constrains the entire system. For example, being at leading on Dimension 1 while Dimension 3 remains absent cannot deliver human-in-the-loop discipline at scale, simply because the available tooling does not support it (§4.3).
3. Three asks
| # | Ask | Type | Owner | Decision by |
|---|---|---|---|---|
| 1 | Name an owner for the institutional benchmarking-and-improvement cycle, mandated to score [INSTITUTION NAME] against Dimensions 1–4 at a published cadence. | Mandate | [NAME] | [DATE] |
| 2 | Approve an AI procurement standard covering the six observable properties (verifiable citation, data residency, uncertainty reporting, reproducibility, auditability, and open-source-and-local options), along with an evaluation gate that applies it before deployment. | Decision | [NAME] | [DATE] |
| 3 | Fund AI literacy as a recurring budget line, including research methods curriculum content, advisor and supervisor development, and examiner training (examiners refers to dissertation committee members in US usage). | Resource | [NAME] | [DATE] |
Table 3. Proposed institutional requests with type classification, assigned owner, and target decision date.
If [COMMITTEE] does not grant them:
- If there is no mandate: Dimension 5 stays at absent. While Dimension 5 remains below established, any established or leading claim in Dimensions 1 to 4 is unassured. Lower documentary levels remain valid descriptions of what the evidence supports (§4.1).
- If there is no procurement standard: Tool choices remain private to each laboratory. We cannot provide evidence regarding where unpublished research data is processed, and our Article 27 assessment lacks any documented artifact base.
- If there is no funding: AI literacy training remains voluntary and uneven across departments. As a result, advisors, supervisors, and dissertation committee members will not be equipped to check what our institutional policy requires when certifying a degree.
4. Regulatory exposure
| Jurisdiction | Obligation | Applies from | Status | Owner |
|---|---|---|---|---|
| European Union | EU AI Act (Regulation (EU) 2024/1689) Article 27 (Fundamental Rights Impact Assessment) for deployers of high-risk systems: admissions, learning-outcome evaluation, education-level assignment within institutions, and proctoring. Applies to institutions operating in or serving the EU market. | 2 August 2026 | [Not started / Under way / Complete and registered] | [NAME] |
| Australia | Privacy Act 1988 (Cth), as amended by the Privacy and Other Legislation Amendment Act 2024, regarding automated-decision-making transparency. Under this requirement, the privacy policy must state the kinds of decisions made by substantially automated means and the personal information used in them. Applies to institutions subject to that regime. | 10 December 2026 | [Not started / Under way / Complete] | [NAME] |
Table 4. Regulatory obligations by jurisdiction with effective dates, current compliance status, and designated owners.
Both of these regulations apply directly to educational and administrative AI systems rather than academic research use itself. For each applicable obligation, a named owner, documented compliance plan, and published timeline supply evidence toward the relevant Dimension 5 criterion (§2.5). One regulatory action does not establish the whole dimension. Dimension 5 reaches established only when its policy, people, systems, and process cells all meet established (§5.1).
5. Why there is no single readiness score
To be read aloud. We are not giving [COMMITTEE] a single overall readiness score, and that is deliberate. Institutional readiness actually comes down to five distinct areas. First, it depends on whether we have clearly defined which research judgments must remain human. Second, it depends on whether that standard is actively applied in advising, supervision, and dissertation committees. Third, it depends on whether the software and tools we procure make that standard easy to follow. Fourth, it depends on whether our researchers, faculty, and students have the necessary AI literacy to do this work. And fifth, it depends on whether we routinely benchmark our progress and publish the results.
Calculating an average score across these areas would allow a high rating in one dimension to hide a serious deficiency in another, even though that deficiency is what actually limits institutional capability. For that reason, we report five separate levels, identify the specific constraint holding us back, and ask for the three concrete decisions outlined above.
Presented by [NAME], [TITLE], [DATE]. Documentary coverage status: ☐ Final ☐ Provisional. D5 assurance: ☐ Assured ☐ Unassured. Final or Provisional concerns A1 documentary coverage only. When D5 is below Established, only Established or Leading claims in D1 to D4 are unassured. Lower documentary levels remain valid descriptions.
Crosswalk: Item to report section
| Item | Report section |
|---|---|
| Profile row D1: Human-in-the-loop discipline | §2.1, Appendix A row 1 |
| Profile row D2: Responsible use in practice | §2.2, Appendix A row 2 |
| Profile row D3: Tooling that promotes responsible use | §2.3, Appendix A row 3 |
| Profile row D4: AI-literate humans | §2.4, Appendix A row 4 |
| Profile row D5: Institutional benchmarking grid | §2.5, Appendix A row 5 |
| Reading rule: Level is the lowest of four axes | §4.1, §2.5 |
| Coverage and the Provisional marking | A1 scoring convention, resting on §4.1 (levels represent observable behavior, not assertions) |
| Since last report column and cell tally | §5.1 (third action), Appendix A row 5 |
| Block 2: Binding dimension and its constraint | §4.3 |
| Ask 1: Name the benchmarking owner and cadence | §5.1 (first action), §2.5 |
| Ask 2: Approve the procurement standard and gate | §2.3, §5.1 (fourth action), Appendix A row 3 |
| Ask 3: Fund literacy as a recurring budget line | §2.4, §5.1 (fourth action), Appendix A row 4 |
| Regulatory exposure: European Union, Article 27 | §2.3, §2.5, §5.1 (second action) |
| Regulatory exposure: Australia, Privacy Act 1988 | §2.5, §5.1 (second action), Appendix A row 5 |
| Block 5: Why there is no single readiness score | §4.3 |
| Preparer note: Class A to D is a separate descriptor | §1.3 |
| Preparer note: Other-jurisdiction pressure points | Appendix F |
| Preparer note: What Dimension 5 leading requires | §4.2, Appendix A row 5 |
Table 5. Mapping of specific assessment items to corresponding report sections and reference locations.
Michael J. Zyphur, PhD · Professor and Director, Instats · instats.org · support@instats.org
Cite the pack. Zyphur, M. J. (2026). The Institutional AI Readiness Pack: Self-Assessment and Implementation Tools for Responsible AI in Academic Research. Instats Policy Series. https://doi.org/10.61700/bv2nulyhht
Companion report. Zyphur, M. J. (2026). Responsible AI in Academic Research: A Competency Framework for Research Training. Instats Policy Series. https://doi.org/10.61700/t31oy23grr
License. The pack and its instruments are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You may adapt them for institutional use with attribution.