Whose future is it anyway? (Part 2: Who gets left behind?)

In the first post, I looked at the idea that AI is pulling us in two directions at once.

Massively accelerated disruption on one side, and the promise of prosperity on the other.

What both framings gloss over is a more uncomfortable question:

Who actually gets access to that prosperity?

Because AI isn't arriving evenly.

AI is not global

There’s a tendency to talk about AI as if it is a global phenomenon, but in practice, it's highly concentrated.

The frontier models are developed by a small number of organisations, mostly based in:

• The United States
• China

They're backed by:

• Enormous capital
• Access to advanced compute
• Deep pools of technical talent
• Strategic government interest

That combination isn't widely distributed, and that matters, because if AI becomes a primary driver of productivity — or possibly the primary driver of global economic productivity — then access to AI capability becomes a fundamental determinant of economic power.

The disappearing ladder

For much of the last few decades, developing economies have had a relatively clear path to growth.

Not an easy one, but a visible one.

Industrialisation → services → integration into global markets.

Labour arbitrage played a significant role in that journey.

Countries could:

• Offer lower-cost labour
• Attract outsourcing
• Build capability over time
• Move up the value chain

AI threatens that model, because it targets exactly the kind of work that formed the next rung of the ladder:

• Back-office processing
• Customer support
• Software services
• Routine analysis
• Data labelling and annotation

If that work is automated before countries can fully build capability around it, then the ladder doesn’t just become harder to climb, it starts to disappear.

When the work disappears

It’s easy to talk about “jobs being automated” in abstract terms.

It’s much harder to think about what that means in practice.

For many economies, particularly in parts of the Global South, large-scale employment has been built around providing relatively low-cost, human-intensive services into global markets.

Call centres are the obvious example, but if we look at the types of jobs I just mentioned it clearly extends much further.

These are not just jobs, they are economic foundations that support:

• Families
• Local economies
• Education
• Upward mobility across generations

If AI automates significant portions of that work, the impact isn't just technological, it's social, economic, and deeply personal.

Displacement without transition

Historically, automation has been disruptive, but it has often been accompanied by transition.

Old jobs disappear, new ones emerge.

People move, adapt, retrain.

Less farriers shoeing horses, more mechanics fixing cars.

That transition is rarely smooth, but it exists.

The concern with AI is that the transition compresses the period in which work disappears — the jobs of today end up disappearing faster than new categories of work for the jobs of tomorrow are created, and faster than people can realistically retrain.

That creates a gap, not just in employment, but in identity, stability, and opportunity.

The risk of deeper exploitation

There is also a darker possibility.

Not that labour disappears entirely, but that it becomes even more commoditised.

If AI handles the high-value work, what remains may be:

• Lower-paid
• More fragmented
• More precarious
• More tightly controlled by platforms

In that scenario, countries that previously benefited from labour arbitrage may not move up the value chain.

They may be pushed further down it.

More dependent. Less empowered.

Governments without good options

This is where the problem becomes political.

Governments are not starting from a neutral position.

They are dealing with:

• Existing inequality
• Limited resources
• Conflicting incentives
• Short electoral cycles
• Global competition

Even under stable conditions, coordinating economic transition at scale is hard.

Under conditions of rapid technological change, it becomes significantly harder.

Particularly when:

• The technology is controlled elsewhere
• The pace of change is externally driven
• The economic model is being disrupted from the outside

There is no obvious playbook for this.

More than a technology problem

This is why framing AI purely as a technology shift misses the point.

For some countries and communities, it represents:

• A loss of economic role
• A challenge to social stability
• A test of political capacity

The problem is not just that some countries move faster, it’s that others may lose the ability to move at all.

The question is not simply whether they can "adopt AI", it’s whether they can absorb the consequences of it.

Leapfrogging, or exclusion?

There is an alternative view, where developing economies may be able to leapfrog.

We’ve seen versions of this before:

• Mobile phones skipping landlines
• Digital payments skipping traditional banking

AI could follow a similar pattern.

Countries without legacy systems may be able to adopt new economic and technological models faster.

Less friction. Fewer constraints.

That sounds pretty plausible, but it still depends on access.

Access to:

• Infrastructure
• Compute
• Education
• Capital
• Platforms

Without that, leapfrogging becomes much harder.

The infrastructure question

This is where things get more geopolitical.

Infrastructure is not neutral — who builds it matters.

China’s investment in digital infrastructure through initiatives like the Belt and Road and the Digital Silk Road is often framed as development support.

It is that, but it's not investment without agenda, and it's building to something much bigger and more impactful in the long term.

Infrastructure defines:

• Standards
• Dependencies
• Influence
• Data flows

Countries that adopt that infrastructure are not just gaining capability, they are also entering into a system.

Before I get accused of being anti-China in particular, it should be noted that the same is true of western governments and western platforms.

Different political and economic models. More extensive and expensive lobbyists. Similar dynamics.

Capability and control

At a global level, the AI question becomes less about technology and more about capability and control.

• Who can build models?
• Who can run them?
• Who can afford to use them at scale?
• Who owns the interfaces through which they are accessed?

These are not evenly distributed, and they are not trending towards being evenly distributed either.

So who gets left behind?

The answer is not fixed, but the risk is clear.

Without deliberate effort, the benefits of AI are unlikely to be anything close to evenly distributed, equitable, or in any way neutral.

They will flow towards:

• Capital
• Capability
• Control

And away from those who lack them.

Which brings us back to the original question.

Not what the future looks like, but:

Who gets to be part of it?

What comes next

In the next post, I’ll look at how this dynamic plays out at the level of companies and platforms, because even within the same country, access is not equal, and what starts as empowerment can very quickly become dependency.