How To Compare Possible AI Futures

A regional council is considering support for AI in businesses, schools and public information services. Its budget is fixed. A presentation assumes that better models, lower costs, abundant power, skilled staff and public trust will arrive together. The council needs to decide which parts of its plan can withstand less convenient conditions. This is Lesson fourteen, How To Compare Possible AI Futures.

We will build two fictional scenarios, trace their consequences and identify evidence that would change the decision. Start by stating the decision and period. In our exercise, the council is choosing how to allocate a fixed public budget over the same evaluation period. It wants useful services and opportunities that people can actually reach. A scenario describes a coherent possible future for testing that choice. A forecast additionally requires a forecasting method

and support for its probability claims. Our scenarios have no supplied probability; their value comes from making the assumptions and consequences inspectable. Choose a few uncertainties that could change the plan. We will examine effective access and skills, infrastructure and external services, and the capacity to check results and respond to people. Keep the headline model capability and the public budget the same across both scenarios. Then ask how those conditions alter the path from a capability to a useful result.

This prevents us from making one future attractive simply by improving everything at once. In Scenario A, technical service costs fall and established firms adopt quickly. Training and accessible support remain limited. Assume existing external computing services can meet this scenario’s initial workload. Firms with complementary skills and finance may redesign their work, while smaller teams and some residents struggle to use the same capability. The council should trace the actual task: what becomes easier, which duties remain and whether the complete service improves.

Adoption by well-equipped firms could create value while leaving participation uneven. Follow that value to workers and customers. A verified reduction in effort might support more service, different roles or a larger margin. Bargaining, ownership, competition and organizational choices affect which path is taken. Staff may need new skills or face a harder mix of remaining work. Scenario A therefore needs evidence about full tasks, pay, hours and opportunities. A faster technical step cannot settle those outcomes by itself.

Schools and information services also receive capable tools in this scenario. Learners can obtain explanations and editors can generate summaries, while teacher review time and checking capacity remain unchanged. The council should examine whether independent learning and dependable public information improve under those conditions. More completed material may increase demands on assessment and verification. These are pathways to investigate within the scenario, rather than observed regional results.

In Scenario B, technical service costs also fall, while the same public budget is allocated differently: shared access, training and usable correction routes receive more support. That uses resources that could have funded other services or equipment. Smaller teams and teachers may gain practical support, but infrastructure delivery remains slow and an external supplier still provides an essential service. Better training leaves those supply constraints to be managed.

This future changes some opportunities while retaining difficult trade-offs. Now place demand inside Scenario B. More supported users may attempt more tasks. Lower effort per task can coexist with greater total work and computing demand. Trained staff, available power or operating capacity may limit what can be delivered. A staged connection or a service interruption can change the feasible timetable. The council should test a viable alternative for an essential function and record which tasks remain constrained,

rather than assume that support makes every expansion possible. The two scenarios now expose different consequences. A worker's opportunity depends on training and role design; a smaller firm needs usable access; a teacher needs conditions that reveal learning; a resident needs information and correction they can rely on. Infrastructure choices carry local costs and affect which services can operate. Ask which assumption changes each person's position and who has influence over the response.

Keep those consequences connected to the decisions that produced them. Documented scenario practice helps clarify the method. The International Energy Agency's World Energy Outlook 2025 explicitly distinguishes its scenarios from forecasts and states their policy starting assumptions. Its October 2026 report on powering AI also examines uncertainty about data-center connections and the timing of operating demand. These dated analyses have their own scope. They show why a plan should retain assumptions

and milestones instead of treating a queue or a published pathway as a guaranteed future. Give each important link an indicator. For work, compare accepted output and complete effort, then inspect roles, hours and training opportunities. For access, observe which intended users can complete the task and obtain correction. For schools, collect independent learning evidence. For information services, examine claim checks and correction responses. For infrastructure, retain confirmed delivery and operating records.

Choose the measures needed for this decision and state which observation would weaken the proposed pathway. The council can now identify actions useful across the two scenarios. A clear acceptance standard, usable source records and a tested correction route may help under different adoption patterns, although each requires staff and resources. Other commitments depend on a specific condition. An expansion relying on a connection date should state the delivery evidence required

and what happens if it is delayed. Compare alternatives and their opportunity costs; a trial is one option to examine, with its suitability depending on the task and stakes. Write the decision with a revision condition. Which action is proposed, which assumptions support it and what observation would trigger expansion, delay or a different approach? Involve the people whose work, access and local resources are affected. The point is to make uncertainty usable in a decision,

while keeping the ability to respond when evidence changes. A scenario that cannot identify a consequence for the choice needs further development. Try the Learning Check. What should every scenario show? Identify the assumptions, relationships, limits, decision and evidence that make it useful for comparison. Pause and write your answer. Every scenario should show assumptions, causal relationships, constraints,

affected decisions and indicators that would support or weaken it. Keep the future coherent and distinguish it from a forecast or an observed result. Explain how a change reaches the decision and which evidence would make you reconsider the pathway. Return to the council's presentation. We can now question its bundled assumptions and compare two futures without declaring either inevitable. We can trace work and value,

participation, resource constraints and institutional choices, then attach a decision and revision condition to the evidence that matters. That gives the council a reasoned basis for choosing and updating its plan. For Practice, use the different fictional regional translation service in your worksheet. Build two scenarios from the supplied uncertainties, trace demand and reviewer capacity, and examine placement and pricing choices. Define indicators and decision triggers. Keep supplied facts separate from assumptions,

without inventing probabilities, market sizes or a successful result. Make your own comparison before using the separate key. This completes the numbered lessons. Our three course questions now connect: how AI changes work and value, how it changes society and who benefits, and who shapes the outcome. A capability reaches wider consequences through tasks, skills, access, resources, organizations and institutions. Use the separate course practice and five-page project to explain those chains,

assess important claims and compare conditional futures. Your account can change as stronger evidence and different choices alter the pathways.