What If — researching implementation to 2035

Purpose and authority

What If research starts from a proposition and investigates how it could be implemented by 2035. Its product is a practical, evidence-informed proposal: the experience people would have, the changes required, who could deliver them, what they would cost, the sequence of action and the decisions that could change that sequence.

This research-family context is dated 9 September 2026. The family builds on ORW's hypothesis-research reader pattern but changes the central question. Hypotheses Research asks what the evidence supports about a claim. What If asks what could responsibly be built from an explicitly stated proposition. Empirical questions inside the design still receive ordinary evidence appraisal.

The first programme is What if Australians had time and personal AI opportunity by 2035? Earlier ORW hypotheses and their verdicts remain unchanged. This guide supplies the reusable public method for this family; unpublished antecedents are not evidence for this release.

1. Write a proposition with a visible boundary

Use this starting form:

What if [people/place] had [opportunity or changed arrangement] by 2035, so that they could [intended public value]? Given that proposition, what would implementation require?

Specify the people affected, the opportunities intended, the freedom they retain, the time horizon and the institutions whose behaviour must change. Separate an attractive future from a claim that it is inevitable. “Could be implemented by 2035” is a research question and planning horizon, not a promise of enactment or success.

Record the proposition's owner and date. Identify assumptions the human decision-maker accepts for design and claims that require evidence. State how the proposition differs from earlier work so inherited restrictions do not silently alter it.

2. Keep the evidence categories separate

An AI-generated result is not a new observation unless it records an actual, appropriately measured event. Agreement between models does not independently confirm their shared assumptions. A demonstration of usability does not establish national employment protection. A public policy announcement does not establish disbursement or benefit.

3. Start with a concrete 2035 experience

Describe a plausible day or decision in an ordinary person's life under the proposition. Mark this as an illustrative scenario. Include people with fewer resources, those who encounter a failure, and those who decline to participate. Make the intended outcome legible without requiring readers to understand a technology architecture.

Then translate that experience into observable requirements. For example, “a person can experiment with AI at no personal cost” implies an accessible account, funded usage, usable services, understandable limits and continuity when a supplier exits. Owning a data centre does not by itself satisfy that requirement.

3a. Compare four implementation pathways

The What If method requires four common pathways: business as usual, two distinct variations, and a fourth pathway in which the proposition succeeds by 2035. Reuse the earlier hypotheses research's discipline of explicit scenarios and common comparisons. The pathway definitions below adapt that discipline to implementation research; they do not rename the older series' pathways or inherit their likelihood estimates.

Use WF1–WF4 consistently across the programme's reports. They are alternative implementation paths, not four compulsory chronological stages. A programme can move directly toward WF4 or use WF3 as a useful interim position. Intermediate pathways must be credible alternatives, not deliberately weak examples chosen to flatter WF4.

Define success before comparison. Specify observable outcomes, who must benefit, the outcome date and what would count as failure. WF4 is a designed success scenario, not evidence that success has already occurred or a forecast that it will occur. Accepting the proposition for design allows this endpoint; the research must still expose every unverified link between an action and that endpoint. If a necessary condition has no feasible route, record the gap and redesign the pathway rather than silently assuming it away. Preserve the current evidence judgement independently of the pathway name.

For every pathway compare the same population, horizon, actual individual experience, intervention scope, recurring and capital costs, accountable roles, delivery timing, employment outcomes, cohesion/harms and evidence gaps. Put common external conditions around all four: slower/faster AI change, constrained public finances, supplier disruption and unequal access. Do not give WF4 a favourable world while testing only BAU against shocks. These stress cases supplement the four pathways; they do not become a fifth pathway.

For each transition toward WF4 record barrier → proposed action → responsible role → resources/authority → checkpoint → evidence of progress → fallback. Carry an unresolved resource or outcome link into the next report. Analyse how WF4 could succeed under stress and where it would cease to satisfy the proposition. Assign no numerical pathway probabilities without a separately justified forecasting model.

4. Use six analytical passes

1. Proposition and person. Define the 2035 outcome, accepted premises, beneficiaries, freedom of choice and scope exclusions. Carry forward relevant previous ORW evidence with its original limits.
2. Precedents and constraints. Examine actual overseas and Australian experience. Record implementation status, recipient, conditions, costs, failures, methods and transfer limits. Include the strongest adverse evidence and credible alternatives.
3. Implementation options. Walk the complete individual journey in WF1–WF4. Compare BAU, incremental reform, partial transformation and the approach that makes the proposition succeed. Explain who does what and who has authority.
4. Resources and consequences. Calculate time, staffing, service capacity and expenditure. Examine distribution, dependency, opportunity cost and possible harm. Vary uncertain inputs; do not manufacture effect sizes or probabilities.
5. Backcast from 2035. Work backwards from WF4's success conditions, identifying what WF2/WF3 change relative to WF1 and what they leave unresolved. Then present an executable forward sequence, with dates, accountable roles, evidence required and actions if milestones fail.
6. Action and review. State what can be recommended now, which human choices remain, and what a bounded next step would produce. Specify revision, pause and stop conditions. Record the handoff to the next report.

The sequence is iterative: evidence may require changing the route, pace, funding mechanism or service offer. It must not be used to conceal a change in the accepted proposition.

5. Backcasting is more than a list of dates

For every milestone record:

Include a route that can deliver useful value before the largest reform is complete. Include a delayed-adoption case and an adverse case. Keep political deadlines separate from implementation milestones: an election may determine when a proposal must be ready without determining when legislation or nationwide service can start.

6. Reuse ORW's reader scaffolding

Every report carries an executive summary, reader orientation, definitions, prominent premise box, numbered passes, evidence/confidence treatment, individual's experience, a WF1–WF4 comparison using the programme's common definitions, economics, responsibilities, distribution and harms, objections, a 2035 backcast from WF4, current design judgement, handoff, references, glossary, authorship disclosure and change log.

Retain a plain confidence scale: high for direct well-supported findings; medium for credible but partial or transfer-dependent findings; low for weak or substantially indirect findings; unresolved where required evidence has not been established. Give confidence to specific claims, not an undifferentiated score for the whole proposition.

Use stable claim identifiers in a shared evidence-and-design ledger. Each row identifies the source or assumption owner, date, location, finding, limitation, affected report and next observation. Report handoffs state what is supported, what is only assumed, and what remains open. Do not sum overlapping participant populations as independent confirmations.

7. The judgement concerns the proposed implementation

Available judgements are:

• DESIGN PLAUSIBLE; DELIVERY UNRESOLVED — enough basis to develop an option, with named delivery gaps.
• READY FOR BOUNDED DEMONSTRATION — the particular demonstration has funding, authority, an accountable operator and a suitable evaluation/safety plan; this does not authorise national rollout.
• REVISE DESIGN — a material constraint changes the current option.
• DEFER PENDING NAMED EVIDENCE — a specific missing input prevents a responsible next decision.
• DO NOT PROCEED WITH THIS DESIGN — the proposed route fails its requirements.

The research can recommend continued design without claiming a field demonstration is ready. Accepted premises do not overrule insolvency, coercion, service failure or unavailable legal authority. If strong evidence directly conflicts with a premise, describe the conflict and consequences for the human decision-maker. Do not suppress it or declare a premise scientifically untestable merely because this family does not require its prior proof.

8. Turn research into an action proposal

The final integrated proposal should identify a minimum useful programme, participant journey, delivery alternatives, proposed governance, capital and recurring costs, staged funding, a 2035 backcast, evaluation design and a concise decision package. It must describe what the One World Community could do within its actual capacity and what requires public institutions or other parties. Do not assume the community already has legal personality, staff or funds.

AI can assist with evidence retrieval, synthesis, accessible explanations, option design, reproducible calculations and proposal assessment. Disclose those roles. People own values, acceptance and final investment decisions. Published material must distinguish researched proposals, simulations and observed programme results.

This family adds a research pattern; it does not rewrite the existing public Hypotheses Research collection. It is labelled What If — implementation research to 2035 and explains the method at entry. Any implementation, funding or field activity remains a separate human decision.

Suggested citation

How this page was derived: the full substantive source body was rendered after the declared public-context transformations, removal of private delivery/workflow material and omission of the private change log.