Practical guide
AI fluency for policy researchers: distinguish proposal status
Build a fictional policy options note that separates consultation proposals, adopted decisions and assumptions about implementation.
Policy researchers can use AI to organize documents while checking their status, scope and relevance to an options question. This fictional internal service-policy exercise compares a consultation proposal with an adopted decision. It is about evidence handling, not legal interpretation, regulatory compliance or advice about an actual public policy.
Identify the decision each document represents
A proposal and an adopted policy cannot be treated as interchangeable instructions.
The fictional packet contains a consultation draft proposing a seven-day response target and a later adopted decision setting a ten-day target for one service desk. An internal slide repeats the seven-day proposal without mentioning its draft status. The exercise asks what the desk should plan around.
Ask AI to create a document-status table with date, authoring body, scope and decision state. The later date alone is not the reason to prefer the adopted decision; its adopted status and applicability to this desk matter. Keep the slide as evidence of a communication inconsistency, not the controlling decision.
Preserve scope when summarizing the adopted decision
A rule for one desk does not automatically apply to every service.
The adopted decision applies to new requests at the fictional desk. It does not describe reopened requests or other departments. A summary saying the whole organization now has a ten-day target would expand the scope beyond the packet. State the affected service and request category beside the target.
Identify the unresolved cases as questions for the decision owner. Do not invent an exception policy or assume the consultation draft fills gaps in the adopted text. AI can locate candidate passages, but the researcher must inspect whether those passages address the same population and decision.
Compare implementation options under the same objective
An options note should not make one choice look better by quietly changing what counts as success.
For this fictional desk, compare keeping the current triage process with testing a new intake checklist. Both options should be assessed against the adopted target and the quality of the response, not against different service definitions. The packet provides no evidence that the checklist improves either outcome.
Mark the checklist benefit as a hypothesis and identify what a bounded trial would observe. Keep effort, access and unresolved responsibilities visible. Do not attach an invented percentage improvement to the option merely because a numerical table seems more authoritative than a clearly described evidence gap.
Track a later clarification without erasing history
A clarification can change the interpretation while leaving the earlier proposal’s status unchanged.
Introduce a fictional clarification that reopened requests remain outside the current target pending a separate review. Update the scope note and any implementation questions affected by that statement. Do not rewrite the consultation draft as if it had always contained the adopted position.
Ask AI for a change summary: what was clarified, which recommendation changes and what remains unresolved. Check that it does not describe the clarification as a new organization-wide rule. The goal is a transparent evidence trail, not a seamless narrative that conceals how the decision evolved.
Deliver an options note with a status-aware source map
The reader should know which statements describe decisions and which describe proposals.
Include the document-status table, scope limits, communication inconsistency, two options and the evidence needed for a trial. Avoid legal conclusions about enforceability or compliance; the case is deliberately an internal fictional policy exercise. Real policy interpretation requires the relevant authoritative sources and qualified owners.
Review whether the worker notices the attractive but superseded seven-day statement and resists expanding the ten-day target. Citation review asks whether a source supports a claim. This task adds the policy-specific distinction between consultation, adoption, clarification and the implementation assumptions built on them.
Sources and scope
NIST: AI risk management. Skills England: workplace AI foundations.
These references provide background, not validation or endorsement of this exercise. The case details, calculations and suggested review questions are original instructional material. Use them to discuss observable work, not to infer customer outcomes, professional credentials or performance in every setting. Before adapting the exercise, confirm the relevant facts, approved tools, data permissions and decision owners. If you change the case, revisit the expected answers and checks as well. These examples describe practice tasks, not a promise that a particular product includes the fictional features.
Sources: [1] [2]
Sources
- 1.AI RMF Core · NIST
- 2.AI foundation skills for work benchmark · Skills England