Job Guarantee or Basic Income? The Debate Over Fighting AI Unemployment

As warnings that artificial intelligence could cause mass unemployment multiply, the job guarantee, a proposal that once sat on the fringes of economic literature, is back on the table. It would have the state offer a fixed-wage job to anyone who wants to work. At the opposite pole of the debate stands universal basic income.

What happened?

A string of warnings from respected figures says AI could cause a major rupture in office work. These warnings are turning an old idea, the state-backed job guarantee, into an option that mainstream politics can discuss. Pieces on the subject cite an observation from the preface to Milton Friedman's Capitalism and Freedom: only a crisis, actual or perceived, produces real change, and the policies adopted when it arrives depend on the ideas lying around at the time.

Why does Friedman's thesis apply now?

Friedman's argument rests on a simple mechanism. Politicians cannot design policy from scratch in the middle of a crisis; they pick from among options that are already available and mature. The rapid deployment of monetary expansion tools in the 2008 financial crisis, and the spread of direct cash transfers within a few weeks during the 2020 pandemic, fit this pattern. Before the pandemic, even the "helicopter money" debate was considered marginal.

Today the crisis has not yet arrived, but a perceived crisis has. Tech companies and banks are shrinking their headcounts, either citing AI or folding it into efficiency drives. It is unclear how much of the layoffs stems from genuine automation and how much from cost pressure and the interest-rate environment. Perception is enough, though: when a company says "a model is doing this role," policymakers have to start preparing. The race between a job guarantee and basic income is therefore a competition of ideas that began before the crisis.

There is also a timing problem. In earlier waves of automation, jobs disappeared and new ones emerged, but the transition took decades. AI aimed at cognitive work could compress that transition. This is a forecast, not a certainty. But thinking in insurance terms, it makes sense to prepare for a scenario that is costly even if its probability is low.

How do a job guarantee and basic income differ?

The two approaches answer the same concern differently. One provides income through work, the other without it.

Advocates of the job guarantee stress that work provides not just income but also routine, status and social connection. They also note that a job guarantee flexes with the business cycle: participation in the program rises when the economy contracts and falls when it recovers. Basic income advocates argue that if most of the value created by AI is concentrated in a handful of companies, a distribution mechanism not tied to employment becomes unavoidable. This argument rests on the assumption that work itself will shrink; the job guarantee instead assumes that work will not run out, but that the market will not offer enough of it.

The weak points of the two models also differ. For basic income, the central debate is financing and whether the payment is adequate: an amount that would truly prevent poverty places a heavy burden on the budget. For the job guarantee, the central problem is finding meaningful, sustainable jobs that match people's skills for millions of people. The capacity of public institutions to design work at this scale remains untested.

What does the comparison with Türkiye's existing programs show?

In Türkiye, İŞKUR's Community Benefit Program (TYP) can be seen as a small-scale, limited relative of the job guarantee. The program places unemployed people in municipalities and public institutions for a fixed period, providing wages and social security support in return. But TYP is not a right: its quotas, duration and eligibility conditions are limited. It is therefore far from the principle of "a job for everyone who wants one" and functions more as a temporary bridge during unemployment.

The most important thing TYP shows institutionally is that a job guarantee is technically feasible. The criticism is that it is temporary and often fails to deliver a lasting transition to paid employment. In a wave of AI-driven layoffs, steering a mid-level banker or a software test specialist toward TYP-type jobs could be a mismatch in both skills and pay. If a job guarantee is to be taken seriously, the program design would need to evolve into a structure that links to the jobseeker's previous profession and is paired with training.

What's next?

There are a few indicators to watch. The first is how much of a role AI really plays in the stated reasons for layoffs; this will become clear in employment statistics rather than in company statements. The second is pilot programs: small-scale job guarantee or cash transfer trials in specific regions produce more valuable evidence than ideological debate. The third is the tax debate; how AI gains are taxed will determine the funding route for both models.

My assessment is this: expecting a single solution is unrealistic. Basic income is a safety net that can be set up quickly, while a job guarantee is a stronger tool against the social cost of long-term unemployment. If the crisis truly arrives, cash transfers will likely come first, followed by selective job guarantee programs running alongside them. If Friedman is right, the real competition is happening today, over which idea is more mature.

For the relevant framing, see Guardian Tech's piece on the debate.