Correction and update — September 11, 2026: This article now identifies the MOSAIC paper’s authors and describes the figures in the publisher-linked January 2026 paper reviewed for this update. The earlier generalized percentile and wage claims have been replaced with explicitly labeled Israeli model scenarios. The original January 29 publication date is unchanged.
If AI and robotics substantially increase productivity, who will receive the gains? Proposals for universal high income begin with that question.
A January 2026 paper by Daniel Schreiber and Niv Shapira at the MOSAIC AI Policy Institute proposes a way to distribute an AI-driven economic surplus. Its title is When AI Takes Our Jobs, It Should Also Pay Our Wages. The framework connects a negative income tax to revenue generated under assumptions about automation, productivity, and employment. It is a policy model, not an implemented universal-high-income program.
From Universal Basic Income to Universal High Income
Universal high income goes beyond a payment intended to cover basic needs. Its ambition is that gains from AI and robotics could support a substantially higher standard of living even when paid work becomes less available. That is a vision whose feasibility depends on both the productivity gains and the institutions used to distribute them.
The MOSAIC framework examines scenarios in which output rises while labor’s share falls. Its authors argue that the key policy problem in those scenarios is capturing enough of the resulting surplus to support households.
The MOSAIC Model: How It Works
The institute describes two core funding channels, supplemented by government savings and program consolidation, with the proceeds distributed through a negative income tax. The components are:
- Multiple funding mechanisms: Combine revenue channels rather than assume a single source can support the whole payment.
- Above-trend revenue: Earmark additional corporate and capital-gains tax receipts associated with AI-driven gains, while keeping those statutory tax rates unchanged.
- Savings (Government Automation Dividend): Use AI to automate government bureaucracy and redirect the cost savings.
- AI-linked Deflation (Captured via Dynamic VAT): As AI drives prices down, adjust VAT rates upward just enough to keep consumer prices stable, capturing the "deflation dividend" for redistribution.
- Negative income tax: Provide a minimum payment that tapers as earnings rise, with preserving work incentives as a design objective.
- Consolidation: Roll overlapping welfare programs into one streamlined payment.
The paper’s Israeli calibration gives an illustrative basic floor of about 6,750 NIS per month per adult-equivalent, or roughly 14,000 NIS for a family of four. Its separate, more ambitious example targets 15,000 NIS per adult-equivalent, approximately 142% of the median wage used in the paper, and requires additional funding mechanisms. These scenario-dependent amounts are not promises of particular income percentiles for households everywhere.
Why Timing Matters
Schreiber and Shapira argue that the political opportunity for adopting such a framework may be greater early in the transition, before disruption and concentrated interests make reform harder. That is the paper’s political-economy argument, rather than an established deadline for action.
Act early, they argue, and we build a foundation for shared prosperity. Delay, and the window may close forever.
Reactions and the Bigger Picture
The proposal has already sparked lively debate. Supporters see it as a pragmatic way to avoid social unrest in a post-work world. Critics worry about disincentives to work or government overreach. Others suggest alternative paths, like treating AI as a partner rather than a resource, or focusing on energy as the new currency.
Regardless of your view, the conversation is shifting. We're no longer just talking about job automation. We're debating how to design the economic rules for a world of radical abundance.
The MOSAIC Model provides a concrete proposal to examine, with published assumptions and replication materials. Whether it could deliver universal high income in practice depends on productivity, revenue capture, policy choices, and implementation. Those conditions deserve as much attention as the headline income figures.
The proposal depends on both increased productivity and institutions capable of sharing its benefits. Each deserves scrutiny before abundance is treated as a certainty.
