MOSAIC

An extreme AI scenario can produce two outcomes at once: mass job displacement and a large increase in productive capacity.

Picture 2050 with 40% unemployment. The familiar fear is total joblessness, starvation and collapse. Yet GDP can still expand if automation creates enough output and the savings are reinvested or paid to shareholders. Abundance does little for social stability when people cannot access it.

Distribution becomes the engineering problem. MOSAIC addresses it without raising tax rates on earned income.

The Automation Paradox

When AI displaces workers en masse, spending need not collapse because the businesses using AI save money. Those savings either get reinvested (R&D, expansion, new ventures) or paid out to shareholders. Either way, money keeps circulating. GDP grows. It grows faster than employment shrinks.

This happened in agriculture. We went from 40% of workers in farming to less than 3%. The economy grew richer. The labour freed up fueled every other sector.

Agriculture changed gradually. AI could compress the same kind of transition into a decade, making speed the problem.

Ring-Fencing Capital Gains

MOSAIC taxes the beneficiaries of automation rather than displaced workers.

When AI companies deploy systems that eliminate jobs, they're capturing a huge chunk of the economic surplus. That's legitimately your business. But it's also an asset extracted from the commons, labour relationships, infrastructure, institutional knowledge, social stability.

Ring-fence those gains through a small, dynamic VAT on transactions where AI is doing work that humans did before. The money goes directly to a distribution mechanism.

The Israel Model

Israel was the first country to seriously model this. Under different Advanced AI (ASI) scenarios, their modeling showed:

• Capital surpluses jump from 420 billion NIS baseline to 1,652 billion NIS under ASI scenarios
• That's 3.93x growth in distributable capital
• They calculated a basic income of 6,751 NIS monthly ($1,800 USD equivalent) with money left over for infrastructure and services

Those figures come from the modelled arithmetic of automating healthcare, customer service, knowledge work and logistics together.

Why early design matters

Waiting allows the beneficiaries to entrench. They build political power. They capture regulation. They convince everyone that what they're doing is "natural market forces" rather than a policy choice.

If you implement ring-fencing after the displacement happens, you're extracting from entrenched winners. That raises the risk of conflict and collapse.

If you implement it while the economic surplus is still visibly growing, the ring fence becomes a design parameter of the system. Companies still win, displaced workers retain spending power and the economy can continue expanding.

The policy is easier to build before the winners become entrenched.

Model it before scaling it

You can model it, run the numbers and test it on small populations. Israel did. Denmark did. Singapore's playing with it.

In the scenarios cited above, unemployment benefits skyrocket because there are too few jobs for the workforce, regardless of people's willingness to work. Basic income provides a floor beneath work income so people do not starve while GDP doubles.

The tax maths works because automation creates an enormous economic surplus. The model distributes part of that abundance.

The political frame

Call it "punishing success" or utopianism and you lose. Present it as a least-bad way to make innovation benefit more people and it might have a chance.

MOSAIC is an unglamorous engineering compromise between markets and socialism, built for a plain economic problem. It can scale, can be implemented and asks the winners to share the gains more deliberately than markets naturally would.

Without a distribution mechanism, the abundance can produce severe instability. Model the mechanism, test it on a small population and make the distribution rule visible before the winners become entrenched.