Preparing for an AI future

A two-hour working session exploring three very different future scenarios. Discuss each scenario and how to respond.

Plateau

Due to technology, economy, or policy, AI progress and use level off.

Diffusion

AI technology makes people more capable, without making them replaceable.

Consolidation

A small set of firms, states, or AIs control access to increasingly effective AI.

Which future do you see?

The world is complex, and we can probably expect to see a blend of different outcomes. Based on the descriptions above, choose a location on the triangle that you think is plausible.

Drag the circle, or click anywhere in the triangle. Preview the host view

All three at once

Where we landed

Where did the group land, and what would move it?

Session structure

These are not predictions, and they are not ranked. They are three worlds to test your plans against. For each one, answer the same four questions. The last one is what you leave with.

Signals How would we know this is happening? List things you could actually watch.

Risks Where could this hurt us? Be specific to this organization, not the economy.

Opportunities Where could we gain? Every one of these futures opens something.

Strategy What do we do, and when do we do it? Each move needs a trigger and an owner.

Plateau

AI progress and adoption level off. This could come from the technology, if capability gains get smaller and harder to win. It could come from economics, if the returns do not justify the spending. It could come from policy, if governments slow deployment down.

Signals

How would we know this is happening?

Examples

Enterprise pilots that don't renew.
New model releases that feel incremental.

Risks

Where could this hurt us?

Examples

Money already spent on tools whose vendors won't survive.
A roadmap that assumes capabilities keep improving.

Opportunities

Where could we gain?

Examples

Cheaper talent and compute.
Winning on reliability while competitors cut back.

Strategy

What do we do, and when do we do it?

Examples

Know which AI bets are load-bearing and which are speculative.
Keep a fallback for anything critical that runs on a single vendor.

Diffusion

AI becomes capable, cheap, and widely available. People across many sectors use it routinely and get more done than they could before. Roles change, but the work still needs people to direct it, judge it, and be accountable for it.

Signals

How would we know this is happening?

Examples

Productivity gains visible in your sector's output, not just in demos.
Competitors shipping noticeably faster.

Risks

Where could this hurt us?

Examples

Falling behind faster adopters.
Security and quality debt from adopting too quickly.

Opportunities

Where could we gain?

Examples

Work that was too expensive to be worth doing becomes cheap.
Moving people off routine work onto things only people can do.

Strategy

What do we do, and when do we do it?

Examples

What to build and what to buy once capability is cheap.
What still sets us apart when everyone has the same tools.

Consolidation

AI keeps getting more capable, and the frontier pulls further and further away. Access to that frontier is important to stay competitive. People may start to play a shrinking role in the workforce, and seek increasing social supports.

Signals

How would we know this is happening?

Examples

Rising cost of switching away from one or two providers.
Decisions that matter being made by systems nobody can overrule.

Risks

Where could this hurt us?

Examples

Margins captured by providers we can't leave.
Exposure to a small number of regulatory or geopolitical chokepoints.

Opportunities

Where could we gain?

Examples

Being the independent option other people can move to.
Making how we govern AI something customers actually value.

Strategy

What do we do, and when do we do it?

Examples

Work out what each provider would cost us to leave.
Name the decisions that stay with a person, in writing, and who that person is.

Synthesis

Now compare the three. The useful answers are the ones that hold up in more than one future.

Moves that work either way

Which strategies above look right in more than one future? Do those regardless.

Signals already showing

Which of the things you listed are happening now?

Owners and next steps

Who watches what, and what happens next week?