Each answer in three lines, with a link to the finding. The answers rest on a stylised simulation model, partly calibrated to Israel, and on descriptive data, so they describe conditions and mechanisms, not forecasts.
“AI will create new occupations, like every earlier technological revolution.”
Not guaranteedThe literature shows that new tasks can be reinstated, but not automatically: in the US over three decades, displacement accelerated and reinstatement weakened. That is why the research deliberately examines a world where it does not happen, as an extreme scenario rather than a forecast.
“Prices will fall, so everyone will gain.”
Only somePeople whose fixed payments are set in shekels do not benefit when the price level falls. When all the saving passes to prices, displaced renters kept their consumption in 11 of 24 model paths, against 17 of 24 for owners without a mortgage. When the saving goes to wages and the price level is stable, the gap disappears: 15 of 24 in both groups.
“We will just tax companies and robots.”
Depends on what can be taxedIn the model, the share of profits within reach of tax is one of the strongest factors: when all profit is reachable, displaced workers kept their consumption in 78% of cases, and when only half is reachable, in 50%. Part of the gap comes from reachability falling gradually with automation in the model. Israel’s corporate tax is about 3.9% of output; from an Israeli base year with a 2% primary deficit, preservation is 57%.
In the model the profit levy is costless. When the tax base erodes and investment responds to the tax, its advantage shrinks but survives up to an erosion elasticity of about 2.4, far above the estimates (0.13–0.58). It is a trade-off, not a free gain: with combined costs in the demand version it comes with more fiscal crises.
Cost of the levy (step 301) · Robustness checks (step 304) · Results chapter →
“A basic income will solve the problem.”
The level yes; the gap depends on financingA benefit that keeps today’s living standard protects against a fall, but not against a gap that opens with those who kept their jobs. In the median case, retained workers ended 29% above the economy without automation, and displaced workers only 3%–4% above it. In 47% of the assumption combinations both groups of displaced workers ended below the economy without automation, even when the government imposes a profit levy before cutting the benefit. The main lever is the benefit’s replacement rate.
“AI is already wiping out jobs.”
Not yet visible in the dataIn 13 European countries there is still no significant link between AI adoption and employment in exposed occupations, and where it is significant it already appears before adoption rose. In Israel, clerical workers were over-represented among the unemployed even before AI tools, and the share of professionals among the unemployed rose from 16% in 2017 to 27% in 2024. Too early to conclude.
“My occupation is exposed to AI, so it will disappear.”
Not necessarilyExposure measures what share of tasks a machine could perform, not whether the job will disappear. According to the authors of the ILO index, a change in the content of work is more likely than its disappearance.