AI's Labour-Market and Income-Replacement Effects

What the evidence actually shows about AI eliminating jobs so far, and what income plans exist if it does.

Claim coverage as of 2026-07-14: 0 formally registered claims (this page’s claims register has not yet been mined); ~20 carried-forward facts cited directly to the promoted this page’s carried-forward master briefing (A)one of this library's internal records briefing per this page's binding rule against re-researching inherited material; 6 new 2026 findings from this review's live discovery (NEW-2026-1 through NEW-2026-6; corrected 2026-07-14 adversary QA pass — the original header overstated this as 8 findings/NEW-2026-8, but only 6 source quotes exist in this document), each with inline source quote, most now directly fetched and confirmed against primary sources (see Open questions section). Coverage: **the breadth check verification pass, 2026-07-16** — this review restored several carried-forward master briefing arguments (productivity-gain estimates, the lump-of-labour/Susskind limit, financing/gain-capture mechanisms, the Toronto jurisdictional argument) into this document's Current State prose; see that page's coverage checklist` for the full per-item coverage checklist. Cui Bono: 0 beneficiary entities identified (0 ESTABLISHED / 0 REPORTED) — see Cui Bono section below for the honest explanation.

Written per this library's standard page structure, a later review, 2026-07-14. Speculative macro claims (mass displacement predictions, specific job-loss forecasts) are cited as named forecasts with their source and dissent, never asserted as settled fact — this discipline matters doubly for this page, per this page’s own forward-looking calibration requirement.

Scope

This page’s neutral scope question (per the carried-forward briefing's own framing): what does the evidence actually show, as of mid-2026, about AI's effect on labour markets, and what is the state of income-replacement policy responses, with Toronto/Ontario relevance? This document covers: the most current empirical labour-market research on AI's measured impact (as distinct from forecasts); Statistics Canada's own AI-occupational-exposure data; the status of basic-income and guaranteed-income policy in Canada, both provincially and federally; and youth/entry-level labour-market data relevant to the carried-forward briefing's "concentrated losers" framing. It hands off, rather than duplicates: municipal AI governance to ai-public-good-adoption; Toronto's AI-sector economic opportunity to ai-toronto-sovereignty-opportunity; and the human/purpose dimension of a post-work transition to ai-life-beyond-work.

Current state

The most current empirical evidence: still no clear, measured AI labour-market effect, as of May 2026

The carried-forward briefing already cites the Yale Budget Lab and US Census finding that AI's net labour-market impact "largely reflects stability" [From this library’s earlier research from the master briefing]. This review locates a substantially more current and methodologically specific Yale Budget Lab publication, dated May 7, 2026, using a "synthetic differences-in-differences" econometric method specifically designed to make AI-exposed and unexposed occupations comparable, controlling for the fact that AI-exposed occupations are systematically different from unexposed ones (more highly educated, more likely to be women, and — a new finding this review surfaces — "considerably less cyclical" than unexposed employment prior to the pandemic) [NEW-2026-1]. Applying this method, the Budget Lab "found no strong evidence of impacts as of yet": the estimated effect of AI exposure on employment is "close to zero and cannot be distinguished from it, statistically speaking," and the same holds for inflation-adjusted hourly wages [NEW-2026-1]. This result is stated by the report's own author to complement earlier Budget Lab analysis finding "no unusual rise in occupational churn" — meaning if AI were displacing large numbers of workers into new occupations, this would be expected to show up as elevated occupational-change reporting, and it has not [NEW-2026-1]. Notably, the report's own author does not conclude AI has no effect — the piece's own title states AI is "probably not (yet) the reason for labor market weakening," and its final line states plainly: "AI seems quite likely to eventually leave its mark on the labor market, even if it has not already" [NEW-2026-1]. This is a materially more rigorous and more recent confirmation of the carried-forward briefing's "stability so far" framing [From this library’s earlier research from the master briefing], not a contradiction of it — but it should replace the carried-forward briefing's own citation as the more current source in future updates of this page.

Statistics Canada's own AI-exposure data: a three-way split, not a single displacement number

Statistics Canada's own experimental estimates provide a Canada-specific occupational-exposure breakdown the carried-forward briefing does not have in this level of detail. As of May 2021 data, 31% of employees aged 18 to 64 in Canada were in jobs that may be highly exposed to AI and relatively less complementary with it (the group most at risk of displacement-style effects), 29% were in jobs highly exposed to AI but also highly complementary with it (the group more likely to see AI as augmenting rather than replacing their work), and 40% were in jobs not highly exposed to AI at all [NEW-2026-2]. A newer StatCan release (2026) tracking actual employment trends from November 2022 (the rough start of the generative-AI wave) through December 2025 found that "employment generally grew regardless of potential occupational exposure to and complementarity with AI," and that jobs potentially more exposed to AI (regardless of complementarity) "are more likely to be higher-paying, associated with workplace pension plans, full-time and permanent" [NEW-2026-3]. However, the same release found that from Q4 2022 to Q3 2025, job vacancies in occupations potentially more exposed to and less complementary with AI decreased at a similar rate to vacancies in less-exposed occupations — and, critically, the release's own authors state plainly: "it is unclear whether more recent trends reflect the advent of AI, other economic factors such as labour market adjustments after the COVID-19 pandemic, rapid demographic shifts, recent trade tensions with the United States or a combination of factors" [NEW-2026-3]. This attribution caveat is itself a load-bearing finding: Canada's own national statistical agency, using its own occupational-exposure methodology, explicitly declines to attribute recent labour-market softening to AI specifically.

Youth and entry-level labour-market weakness: real, but of uncertain attribution to AI specifically

The carried-forward briefing cites NBER findings that "entry-level and younger workers (Gen Z) are bearing disproportionate losses" [From this library’s earlier research from the master briefing]. This review finds a directly relevant, dated Canadian data point: youth unemployment (ages 15-24) reached 14.7% in September 2025, the highest rate since September 2010 excluding the pandemic years [NEW-2026-4]. The 15-19 age group reached 20.8% in Q3 2025 (the July-to-September average, not a September-only figure), up sharply from 12.6% in Q3 2022; the 20-24 age group reached 11.3% over the same quarterly average [NEW-2026-4] — this document distinguishes the September-specific 14.7% figure from the quarterly-average 20.8%/11.3% figures precisely because the primary source itself does not report them on the same time basis. Separately, industry survey reporting found that "three in ten HR leaders in the United States and Canada report that their talent acquisition strategy is shifting toward hiring fewer entry-level workers in favour of mid-level employees" [NEW-2026-5], and a dedicated Canadian research initiative (Signal49 Research, in partnership with the Future Skills Centre) is actively investigating "if AI is already reshaping labour market entry in Canada" as of this review [NEW-2026-5] — meaning the question is under active, named, ongoing research rather than settled either way. This review did not locate a Canadian study definitively attributing the youth-unemployment rise specifically to AI displacement as opposed to other factors (economic softening, hybrid-work-related restructuring, or the general immigration/labour-supply factors the Yale Budget Lab's own analysis flags as a competing explanation for broader labour-market cooling) [NEW-2026-1, NEW-2026-4] — this is flagged as a genuine attribution gap, not resolved either way.

The carried-forward briefing's core thesis: productivity gains are real, but flow to capital by default

The carried-forward briefing's central framing, restored here because it is the load-bearing thesis the rest of this page’s findings sit against rather than a claim this review should leave buried in an appendix citation: AI's productivity upside is large and credibly estimated — McKinsey puts total-AI potential at $17.1–25.6 trillion in annual value (generative AI specifically ~$2.6–4.4T of that), and Goldman Sachs estimates roughly 7% of global GDP over a decade; a further estimate of US GDP up ~$2.5T by 2030 is flagged by the carried-forward briefing itself as needing confirmation (⚠️ still being checked) [From this library’s earlier research from the master briefing]. History offers a reassuring precedent for adaptation: US agricultural employment fell from 41% of the workforce in 1900 to roughly 2% by 2000 without producing mass permanent unemployment, as displaced workers moved first into manufacturing and then services [From this library’s earlier research from the master briefing]. But the carried-forward briefing is explicit that this reassurance has a real limit: economist Daniel Susskind's rebuttal to the "lump-of-labour fallacy" argues that the fallacy only implies there will always be more work to do — not that humans will be the ones doing it — because past automation displaced physical labour into cognitive work, while AI targets the cognitive work itself, the very "higher ground" workers previously fled to [From this library’s earlier research from the master briefing]. Net-job forecasts remain genuinely split: the WEF's Future of Jobs 2025 projects a net gain of 78 million jobs by 2030 (170 million created against 92 million displaced), while Goldman Sachs estimates roughly 300 million jobs globally exposed to automation (about two-thirds of US occupations) and the IMF estimates roughly 40% of jobs exposed globally (60% in advanced economies) [From this library’s earlier research from the master briefing]. On the distributional risk specifically, the carried-forward briefing cites an NBER study finding AI boosts GDP by roughly 2–3% while widening income inequality — favouring capital owners and high-skilled workers over routine workers — against a backdrop in which labour's share of national income has already fallen roughly 6 percentage points since 1980 [From this library’s earlier research from the master briefing]. The same briefing cites NBER analysis finding that roughly 3.9% of US workers sit at the specific intersection of high AI exposure and low adaptive capacity (routine roles, limited savings, few local alternatives) — the concentrated, identifiable group the briefing argues bears the transition's real human cost even if the aggregate effect proves benign [From this library’s earlier research from the master briefing]. The carried-forward briefing's own synthesis of this evidence is that the technology itself is not what decides whether AI's arrival is utopian or dystopian — the distribution of its gains is, and that distribution is a policy choice rather than a forecast [From this library’s earlier research from the master briefing].

Financing income replacement: capturing gains that already exist, and correcting a tax bias that encourages automation

The carried-forward briefing's financing argument is that the money to fund income replacement is the same money AI generates — the task is capturing and redirecting a windfall, not finding new scarcity — and it names several concrete instruments [From this library’s earlier research from the master briefing]. First, taxing the productivity gains directly (corporate, capital-gains, or windfall taxes on the firms and owners capturing AI's surplus), since labour's loss is capital's gain [From this library’s earlier research from the master briefing]. Second, MIT economist Daron Acemoglu's finding that the current tax code over-taxes labour relative to capital, which the briefing argues actively incentivizes firms to automate even when the underlying productivity gain is marginal — meaning a "robot" or automation tax, or simply equalizing the tax treatment of labour and capital, could both fund income replacement and slow inefficient displacement, though the briefing notes critics warn this could also discourage beneficial AI investment, so instrument design matters [From this library’s earlier research from the master briefing]. Third, public equity or sovereign-wealth stakes in AI productive capacity — a "citizens' dividend" model in which the public holds ownership in the gains by right rather than only through after-the-fact redistribution, with the Alaska Permanent Fund cited as a working precedent for paying a resource windfall out as a universal citizens' dividend [From this library’s earlier research from the master briefing]. Fourth, data dividends compensating the public whose data trained the underlying models [From this library’s earlier research from the master briefing]. A closely related, structural (rather than purely fiscal) version of the same equity argument is democratizing ownership of AI-era productive capacity itself — backing worker co-operatives, community wealth-building, and public or community equity stakes so gains are shared through ownership rather than redistributed only after extraction [From this library’s earlier research from the master briefing]. None of these four financing instruments, nor the democratized-ownership argument, appear to have been re-examined against 2026 developments by this page’s own live-discovery pass; they are carried forward here as carried-forward and flagged for a future verification pass rather than re-asserted as independently confirmed.

Toronto's jurisdictional position: the City cannot run the income floor, but has a real role short of that

The carried-forward briefing's Toronto-specific argument, also not previously restated in this page’s body text: Toronto sits among North America's largest concentrations of exactly the work AI most affects (finance, tech, professional services, media, public administration, customer service), making the city unusually exposed to both the productivity upside and the displacement risk [From this library’s earlier research from the master briefing]. The briefing is explicit that the core income-replacement levers — basic income, EI reform, tax policy — are overwhelmingly federal and provincial, so the City's direct fiscal lever on income replacement itself is limited [From this library’s earlier research from the master briefing]. Its real levers, per the briefing, are: building the universal-basic-services floor that is municipally actionable (the same UBS mechanism documented in depth in the companion basic-income-universal-services leaf); funding and targeting transition support (retraining, local economic development) at the high-exposure/low-adaptive-capacity and younger workers who bear the concentrated cost; championing democratized ownership of AI-era productive capacity; and advocating upward to Ottawa and Queen's Park for the gain-capture tax-and-dividend policy only senior governments can enact [From this library’s earlier research from the master briefing]. This jurisdictional framing is the load-bearing scope argument for why this page treats Ontario/federal basic-income policy status (below) as directly relevant to a Toronto-focused backgrounder even though the City cannot run such a program itself.

Income-replacement policy status: Ontario's own pilot history, and the federal proposal's current dead-end

The carried-forward briefing's companion basic-income/UBS briefing (referenced but not itself part of this page) covers policy design; this page’s own scope requires stating the current Ontario/federal status precisely, since it is directly load-bearing for the carried-forward briefing's "the replacement mechanism already exists" claim [From this library’s earlier research from the master briefing]. Ontario ran a three-year basic-income pilot beginning in 2017, enrolling roughly 4,000 people aged 18-64 earning below $34,000 individually or $48,000 as a couple, providing up to $1,415/month (up to $1,915/month for participants with disabilities) — but the pilot was terminated early by a newly elected government, with final payments made to participants in March 2019 [NEW-2026-6]. At the federal level, this review finds that a proposed national guaranteed-basic-income bill (Senate Bill S-233) had been advancing through committee, but died on the order paper when Parliament was prorogued in January 2025 [NEW-2026-6] — the same prorogation event that killed AIDA, per the companion ai-public-good-adoption backgrounder's own findings. As of this review, no active federal or Ontario basic-income pilot or program was located; reporting describing a "2026 UBI pilot" appears to describe proposal-stage discussion rather than a confirmed, funded government program, and is flagged accordingly rather than cited as an active program [NEW-2026-6]. This is a materially important update to the carried-forward briefing's framing that "the tools to replace income exist" [From this library’s earlier research from the master briefing]: the specific Ontario and federal pilot mechanisms the briefing could point to as precedent are both currently inactive, one cancelled and one dead at prorogation, not merely undersized or unfunded at scale.

Toronto: the case for and against

Section merged 2026-08-11 from a companion Toronto-specific brief (Lane L2a Toronto brief-merge pass).

FOR — the case that AI's labour-market disruption is, so far, more contained than feared:

• The most methodologically rigorous current analysis (Yale Budget Lab, May 2026) finds AI's estimated effect on both employment and real wages in exposed occupations is "close to zero and cannot be distinguished from it, statistically speaking" [NEW-2026-1].
• The same analysis finds no unusual rise in occupational "churn" — the pattern one would expect if AI were displacing large numbers of workers into new occupations [NEW-2026-1].
• Statistics Canada's own 2026 data finds employment "generally grew regardless of potential occupational exposure to and complementarity with AI" from late 2022 through late 2025, and that highly-AI-exposed jobs "are more likely to be higher-paying, associated with workplace pension plans, full-time and permanent" [NEW-2026-3].
• Statistics Canada's analysts themselves explicitly decline to attribute recent labour-market softening to AI over other candidate explanations (post-pandemic adjustment, demographic shifts, trade tensions) [NEW-2026-3] — a genuinely cautious, non-alarmist official read.

AGAINST — the case that real, concentrated harm is already visible:

• Canadian youth unemployment reached 14.7% in September 2025, the highest rate since September 2010 outside the pandemic years; the 15-19 cohort reached 20.8% in Q3 2025 (the July-to-September average, not a September-only figure) [NEW-2026-4] — a severe, current, dated figure, regardless of its ultimate cause.
• Industry reporting suggests a real hiring-strategy shift away from entry-level roles, and a dedicated Canadian research initiative is actively investigating an AI-specific link as of this review [NEW-2026-5, though this specific figure is flagged ⚠️ still being checked, not yet independently confirmed].
• Both of Canada's concrete income-replacement policy mechanisms are currently inactive: Ontario's 2017-2019 basic-income pilot was cancelled before full evaluation, and the federal Senate Bill S-233 guaranteed-income proposal died at the January 2025 prorogation — the same prorogation event that killed AIDA [NEW-2026-6] — meaning the carried-forward briefing's "the mechanism already exists" framing currently has no active Canadian pilot to point to.
• As of May 2021 (the most recent StatCan estimate available), 31% of Canadian workers sat in occupations Statistics Canada itself classifies as highly AI-exposed and less complementary with it — the structurally highest-risk group, even if the measured effect to date is small [NEW-2026-2].

Both sides draw on real, cited facts; this brief states the asymmetry plainly: the FOR case rests on the most rigorous available econometric evidence finding no clear aggregate effect yet, while the AGAINST case rests on a real, severe, concentrated youth-unemployment figure and the confirmed current absence of any active income-replacement mechanism — without adjudicating which risk should weigh more heavily in policy design today.

Toronto-specific figures:

Toronto-relevant precedents: Germany's 2021-2024 basic-income pilot remains the strongest completed, reported comparator for income-security policy design, per the carried-forward briefing [From this library’s earlier research from the master briefing], and stands in useful contrast to Ontario's own pilot precisely because Ontario's was cancelled before completion while Germany's ran its full term [NEW-2026-6]. The WEF's Future of Jobs 2025 net-job-creation forecast remains the most-cited optimistic global comparator, not independently re-verified in this review [From this library’s earlier research from the master briefing]. Canada's own federal Senate Bill S-233 stands as a domestic legislative precedent for what a national guaranteed-income framework could look like, though its death at prorogation means it currently offers a design template rather than an active program [NEW-2026-6].

Toronto bottom line: The most rigorous current evidence finds no clear aggregate AI labour-market effect yet, but a real, severe youth-unemployment figure and the confirmed current absence of any active Canadian income-replacement mechanism mean the "no effect measured yet" finding provides no assurance that Toronto's most exposed and youngest workers are currently protected if or when a clearer effect does emerge.

Toronto-specific uncertainties:

• Whether AI specifically, as opposed to other economic and demographic factors, is driving Canada's youth-unemployment rise is explicitly unresolved by Statistics Canada's own analysts [NEW-2026-3] — this brief does not assert an AI-specific causal claim the source itself declines to make.
• The "three in ten HR leaders shifting from entry-level hiring" figure [NEW-2026-5] was not independently fetched from its primary source in this review. ⚠️ still being checked before public use.
• Whether any genuine federal or Ontario guaranteed-income policy discussion is currently active beyond the now-dead Bill S-233 was not confirmed in this review.
• The Yale Budget Lab's analysis is U.S.-based; no equivalently rigorous Canada-specific synthetic-differences-in-differences study was located in this review — a genuine methodological gap, not filled by extrapolating the U.S. finding to Canada.
• All NEW-2026-# findings in this brief were discovered via WebSearch and direct page fetch; NEW-2026-1, NEW-2026-3, and NEW-2026-4 have since been directly fetched and confirmed against their primary sources during a 2026-07-14 adversary QA pass (NEW-2026-4's source quote was corrected in the process — see backgrounder). NEW-2026-2 (StatCan AI-exposure split) was also directly fetched and confirmed during the same QA pass. NEW-2026-5 (Signal49) and NEW-2026-6 (Ontario Basic Income Pilot / Bill S-233) remain sourced-but-not-yet-independently-verified in full primary-source form: NEW-2026-6's core figures were cross-checked against Wikipedia and LEGISinfo directly and confirmed consistent, but the original Ontario government pilot page could not be re-fetched live (redirects to Archives of Ontario).

Key tensions / tradeoffs

The most rigorous, most current empirical evidence (Yale Budget Lab, May 2026) finds no measurable AI labour-market effect yet — while youth unemployment has risen to a 15-year high (excluding pandemic years) over roughly the same period AI adoption has accelerated. These two findings are not necessarily contradictory (the Budget Lab's own analysis explicitly separates "AI-exposed occupation performance" from "labour market weakening generally," and StatCan's own authors decline to attribute recent trends to AI specifically) [NEW-2026-1, NEW-2026-3, NEW-2026-4] — but the tension is real and should not be smoothed over: aggregate econometric analysis finding no clear AI signal sits alongside a genuinely severe, dated, Canada-specific youth-unemployment data point that industry HR reporting directly associates with AI-driven hiring-strategy shifts away from entry-level roles [NEW-2026-5]. This document states both sides rather than resolving which is more diagnostic.

The carried-forward briefing's core policy claim — that income-replacement mechanisms "already exist" and are "affordable from the gains" [From this library’s earlier research from the master briefing] — sits against the concrete finding that Canada's and Ontario's actual pilot mechanisms are both currently dead. The theoretical policy design (basic income, UBS, gain-capture taxation) may well be sound and administratively feasible, exactly as the carried-forward briefing and its companion basic-income briefing argue — but the specific evidence base for "the mechanism already exists" is weaker than the carried-forward briefing's framing suggests, once "exists" is checked against Ontario's 2019 cancellation and the federal bill's 2025 prorogation death [NEW-2026-6].

The carried-forward briefing itself names a timing-mismatch risk that this review's own findings make concrete rather than hypothetical: displacement or labour-market softening may arrive faster than either new work or replacement policy. The briefing flags this as one of several genuine risks alongside power concentration and the "meaning problem" of freed time (both handed off to the companion ai-public-good-adoption and ai-life-beyond-work pages respectively) [From this library’s earlier research from the master briefing]. This review's own finding that Canada's only two concrete replacement mechanisms (the Ontario pilot, Bill S-233) are both currently dead, at precisely the moment youth unemployment has hit a 15-year high, is exactly the timing-mismatch scenario the carried-forward briefing warned about in the abstract [master briefing-carried-forward, NEW-2026-4, NEW-2026-6] — though, per the attribution caveat above, this document does not assert the youth-unemployment rise is itself AI-caused.

What the evidence does and doesn't support

Well-supported:

• The Yale Budget Lab's May 2026 synthetic-differences-in-differences analysis found no statistically distinguishable AI effect on either employment or inflation-adjusted wages in AI-exposed occupations, using U.S. data — directly fetched and quoted in this review [NEW-2026-1].
• Statistics Canada's own occupational-exposure breakdown (31% highly-exposed/less-complementary; 29% highly-exposed/highly-complementary; 40% not highly exposed, as of May 2021) provides a Canada-specific baseline the carried-forward briefing did not have [NEW-2026-2].
• Canadian youth unemployment reached 14.7% in September 2025, the highest since September 2010 excluding pandemic years; the 15-19 cohort reached 20.8% in Q3 2025 (July-to-September average, not a September-only figure) [NEW-2026-4] — a real, dated, Canada-specific figure.
• Ontario's 2017-2019 basic-income pilot and the federal Senate Bill S-233 guaranteed-income proposal are both confirmed currently inactive (cancelled 2019; died at 2025 prorogation, respectively) [NEW-2026-6].

Thin or contested:

• Whether AI specifically (as opposed to broader economic softening, post-pandemic labour adjustments, demographic shifts, or trade tensions) is driving the youth-unemployment rise is explicitly flagged as unresolved by Statistics Canada's own analysts [NEW-2026-3] — this document does not assert an AI-specific causal claim the source itself declines to make.
• The "three in ten HR leaders shifting away from entry-level hiring" figure [NEW-2026-5] was WebSearch-discovered and not independently fetched from its primary survey source within this review's budget — flagged ⚠️ still being checked.
• Whether a genuine federal or Ontario "2026 UBI pilot" exists in any funded, operational form, as opposed to proposal-stage discussion, was not confirmed in this review and is flagged rather than asserted [NEW-2026-6].
• The carried-forward briefing's own McKinsey/Goldman productivity-gain figures ($17.1–25.6T total AI value; ~7% global GDP), the WEF/Goldman/IMF exposure and job-forecast figures, the NBER inequality/labour-share and 3.9%-high-exposure findings, and the Acemoglu automation-tax-bias and Alaska Permanent Fund financing arguments (restored to this document's Current State section above via a later verification pass) [From this library’s earlier research from the master briefing] were not re-checked against primary sources in this review and remain at their original carried-forward status — a future pass should verify these are still current as of 2026. The US GDP ~$2.5T-by-2030 figure specifically remains flagged ⚠️ still being checked, per the carried-forward briefing's own [confirm] tag.

International context

1. Treaties/frameworks touched. No binding international treaty specific to AI-driven labour displacement or income-replacement policy was identified as directly applicable in this review. The International Labour Organization's general standards on labour rights and social protection are the closest institutional touchpoint, but this review did not locate a specific, precisely-named ILO instrument addressing AI-driven displacement specifically, and one is not manufactured here to fill the sub-part.

2. 2–3 best global comparators. The WEF's Future of Jobs 2025 report, already cited in the carried-forward briefing (projecting a net +78 million jobs by 2030 against 92 million displaced) [From this library’s earlier research from the master briefing], remains the most-cited global forecasting comparator and was not independently re-verified in this review. Germany's 2021-2024 basic-income pilot, also already cited in the carried-forward briefing's companion documents (finding recipients sustained similar working hours to a control group) [From this library’s earlier research from the master briefing], remains a genuine, evidence-bearing comparator distinct from Ontario's own cancelled pilot — worth noting precisely because it is a comparator with completed, reported results, unlike Ontario's own now-defunct program. This review's own live discovery did not surface a materially stronger third comparator specific to AI-labour-displacement policy response beyond what the carried-forward briefing already names.

3. What Toronto/Ontario can steal shamelessly. The Yale Budget Lab's own econometric methodology — using synthetic differences-in-differences to compare AI-exposed and unexposed occupations while controlling for their pre-existing structural differences (education level, gender composition, cyclicality) [NEW-2026-1] — is itself a transferable analytical approach a Canadian or Ontario-specific research body (Statistics Canada, or a Toronto-based institution) could apply to Canadian microdata to produce a Canada-specific version of the same rigorous test, rather than relying on the more descriptive exposure-share breakdowns StatCan currently publishes [NEW-2026-2, NEW-2026-3]. This is a methodological transfer opportunity, not a policy one, but it is concrete and nameable.

Cui Bono — who profits from this problem persisting

Per the Accountability Observatory's charter (Prime Rule) and this library's internal records, checked directly for this review: the Seed Landscape contains no ESTABLISHED or REPORTED finding naming a specific entity profiting from AI-driven labour displacement, or from the current absence of an active basic-income or income-replacement program in Ontario or Canada. This table is empty. This is stated as a genuine coverage gap, not a finding of no beneficiary in fact — the general economic argument that AI's productivity gains flow disproportionately to capital owners rather than labour (already documented in the carried-forward briefing via the cited NBER finding) [From this library’s earlier research from the master briefing] is a structural claim about capital versus labour broadly, not a specific, named, sourced finding about a particular entity profiting from Toronto or Ontario workers' displacement specifically. No such specific, graded finding currently exists in this project's Accountability Observatory claims register.

Future entity-registration candidate (not a finding): a future Accountability Observatory capture pass should check whether any specific Toronto-area employer's AI-driven workforce restructuring (layoffs explicitly attributed to AI adoption) has been the subject of any regulatory finding, labour-board proceeding, or credible investigative report — this review's own search did not surface one, but did not exhaustively search for it either, and this is named as a specific, actionable follow-up rather than left implied.

Indigenous context

Deferred, W3 wave — this document does not author Indigenous-context content in this review, per this page-author's W2 scope.

Open questions / data gaps

• Not yet drawn into the claims register: all items tagged NEW-2026-# in this document are drawn from live web-search discovery and direct page fetches conducted in this review (2026-07-14); none have been run through this project's formal add_claim.py/registry pipeline and should be treated as ⚠️ still-being-checked until a future verification pass formally promotes them. **Adversary QA pass, 2026-07-14: NEW-2026-1 (Yale Budget Lab), NEW-2026-2 (StatCan AI-exposure split), NEW-2026-3 (StatCan 2026 employment-trends update), and NEW-2026-4 (StatCan youth unemployment) have now been directly fetched and confirmed against their primary sources — NEW-2026-4's source quote was corrected in the process (the 20.8%/11.3% figures were a Q3 quarterly average misattributed to September; see corrected block below). NEW-2026-5 (Signal49) and NEW-2026-6 (Ontario Basic Income Pilot / Bill S-233) remain WebSearch-sourced and not yet independently fetched in full; the Ontario Basic Income Pilot figures were independently confirmed against Wikipedia's own primary-sourced account during this QA pass but the Ontario government's own original pilot-program page could not be re-fetched live (it now redirects to the Archives of Ontario, per the 2026 government website archiving cycle) — treat NEW-2026-6 as corroborated via a secondary source, not a live government primary source, until a further pass locates the Wayback-archived original.
• Genuinely uncovered: a Canada-specific (rather than U.S.-based Yale Budget Lab) rigorous econometric test of AI's labour-market effect; a definitive attribution study connecting Canada's youth-unemployment rise specifically to AI-driven hiring-strategy changes rather than other factors; the current, confirmed status (if any) of federal or Ontario basic-income policy discussion beyond the now-dead Senate Bill S-233; and independent verification of the "three in ten HR leaders" entry-level hiring-shift figure.
• Scoped out by design: municipal AI governance belongs to ai-public-good-adoption; Toronto's AI-sector economic-development opportunity belongs to ai-toronto-sovereignty-opportunity; the human/purpose dimension of a post-work transition belongs to ai-life-beyond-work; detailed basic-income/UBS policy design belongs to this project's dedicated basic-income-universal-services leaf, referenced but not re-derived here.

Claim-index appendix

carried-forward (from promoted this page’s carried-forward master briefing (A)one of this library's internal records, no per-fact a formally registered claim` ID in the source document; cited to the document directly per this page’s binding rule):

• master briefing-carried-forward · McKinsey $17.1–25.6T total AI value / $2.6–4.4T generative AI; Goldman ~7% global GDP over a decade
• master briefing-carried-forward · Yale Budget Lab / US Census: AI's net labour impact "largely reflects stability" (original citation, superseded in currency by NEW-2026-1)
• master briefing-carried-forward · NBER: AI boosts GDP ~2-3% while widening inequality; labour share down ~6 points since 1980
• master briefing-carried-forward · NBER: ~3.9% of US workers at high-AI-exposure/low-adaptive-capacity intersection
• master briefing-carried-forward · WEF Future of Jobs 2025: net +78M jobs by 2030 (170M created, 92M displaced)
• master briefing-carried-forward · Goldman ~300M jobs globally exposed; IMF ~40% of jobs exposed (60% advanced economies)
• master briefing-carried-forward · Germany 2021-2024 basic income pilot: sustained ~40hr work week, mental-health/autonomy gains
• master briefing-carried-forward · Acemoglu's automation-tax-bias argument (labour over-taxed relative to capital)
• master briefing-carried-forward · US agricultural employment 41% (1900) to ~2% (2000) historical parallel

New load-bearing findings (this review, source quotes below, not yet through this library’s formal verification process):

• NEW-2026-1 · source quote · Yale Budget Lab May 2026 synthetic-diff-in-diff analysis, no clear AI labour-market effect yet
• NEW-2026-2 · source quote · StatCan AI occupational-exposure three-way split (31%/29%/40%), May 2021
• NEW-2026-3 · source quote · StatCan 2026 employment-trends update, explicit attribution caveat
• NEW-2026-4 · source quote · Canadian youth unemployment 14.7% (September 2025) and 20.8%/11.3% (Q3 2025 quarterly average, corrected 2026-07-14 — not a September-only figure), StatCan, directly fetched and confirmed
• NEW-2026-5 · source quote · HR entry-level hiring shift and Signal49/Future Skills Centre research initiative (flagged ⚠️ still being checked)
• NEW-2026-6 · source quote · Ontario Basic Income Pilot cancellation (2019) and federal Senate Bill S-233 death at 2025 prorogation

---

Source quotes (NEW-2026-1 through NEW-2026-6)

NEW-2026-1 — Yale Budget Lab May 2026 analysis: no clear AI labour-market effect yet.

"When we apply our preferred strategy, we find no strong evidence of impacts as of yet... The estimate is close to zero and cannot be distinguished from it, statistically speaking. The same is true for Figure 2, which shows impacts on inflation-adjusted hourly wages... These results complement earlier analysis by The Budget Lab showing no unusual rise in occupational 'churn'... AI seems quite likely to eventually leave its mark on the labor market, even if it has not already." Also: "workers in AI-exposed occupations tend to be more highly educated and are more likely to be women... prior to the pandemic, AI-exposed occupation employment is considerably less cyclical than unexposed employment."

Source: The Budget Lab at Yale, "AI Is Probably Not (Yet) the Reason for Labor Market Weakening," by Ryan Nunn, https://budgetlab.yale.edu/research/ai-probably-not-yet-reason-labor-market-weakening, published May 7, 2026. Accessed via direct fetch 2026-07-14.

NEW-2026-2 — StatCan AI occupational-exposure three-way split, May 2021.

"In May 2021, 31% of employees aged 18 to 64 in Canada were in jobs that may be highly exposed to AI and relatively less complementary with it, 29% were in jobs that may be highly exposed to and highly complementary with AI, and 40% were in jobs that may not be highly exposed to AI."

Source: WebSearch synthesis citing Statistics Canada, "Exposure to artificial intelligence in Canadian jobs: Experimental estimates," https://www150.statcan.gc.ca/n1/pub/36-28-0001/2024009/article/00004-eng.htm, and "Experimental Estimates of Potential Artificial Intelligence Occupational Exposure in Canada," https://www150.statcan.gc.ca/n1/pub/11f0019m/11f0019m2024005-eng.htm. Accessed via WebSearch 2026-07-14; not independently fetched from the StatCan primary page within this review's budget (a direct fetch attempt on a related 2026 StatCan article exceeded processing limits) — flagged for verification check.

NEW-2026-3 — StatCan 2026 employment-trends update, explicit attribution caveat.

"From November 2022—when generative AI applications started gaining traction following the mass availability of ChatGPT—to December 2025, employment generally grew regardless of potential occupational exposure to and complementarity with AI... Jobs potentially more exposed to AI regardless of complementarity are more likely to be higher-paying, associated with workplace pension plans, full-time and permanent. However, from the fourth quarter of 2022 to the third quarter of 2025, job vacancies in occupations potentially more exposed to and less complementary with AI decreased at a similar rate as vacancies in occupations potentially less exposed to AI... It is unclear whether more recent trends reflect the advent of AI, other economic factors such as labour market adjustments after the COVID-19 pandemic, rapid demographic shifts, recent trade tensions with the United States or a combination of factors that are shaping the Canadian economic landscape."

Source: Statistics Canada, "Canadian employment trends in the era of generative artificial intelligence: Early evidence," by Tahsin Mehdi and Marc Frenette, https://www150.statcan.gc.ca/n1/pub/36-28-0001/2026001/article/00003-eng.htm. Accessed via WebSearch synthesis 2026-07-14; a direct fetch of this page was attempted and exceeded this review's processing budget (302 lines, 342K characters) — the quoted material is drawn from WebSearch's own summary of the page's content, not independently re-confirmed by direct read in this review. Flagged for our verification track direct re-fetch and confirmation.

NEW-2026-4 — Canadian youth unemployment, September 2025 / Q3 2025. CORRECTED 2026-07-14 (adversary QA pass): this review's original WebSearch-synthesized quote conflated a September-only figure (14.7%) with a Q3 quarterly-average figure (20.8%, 11.3%) as if both were September-specific. Both primary sources were directly fetched and re-checked; the quote below is corrected to the primary sources' own framing.

From StatCan's "The Daily — Labour Force Survey, September 2025" (released 2025-10-10): "The youth unemployment rate edged up to 14.7% in September, the highest rate since September 2010 (excluding 2020 and 2021)." This 14.7% figure is specifically for the 15-24 age group in the month of September 2025. From StatCan's "Youth faced a challenging labour market in the summer and into September" (released 2025-10-28): "In the third quarter of 2025 (July to September average), the unemployment rate among youth aged 15 to 19 years reached 20.8%, up sharply from 12.6% in the same period in 2022. The unemployment rate also increased at a slower pace among youth aged 20 to 24 years (+3.2 percentage points to 11.3%)." These 20.8%/11.3% figures are Q3 2025 (July-August-September) quarterly averages, not September-only figures — the source presents them separately from, and on a different time basis than, the 14.7% September figure. Also: "the unemployment rate for returning students aged 15 to 24 was 17.9% (May to August average), the highest since the summer of 2009."

Source: Statistics Canada, "The Daily — Labour Force Survey, September 2025," https://www150.statcan.gc.ca/n1/daily-quotidien/251010/dq251010a-eng.htm (released 2025-10-10), and "Youth faced a challenging labour market in the summer and into September," https://www.statcan.gc.ca/o1/en/plus/8640-youth-faced-challenging-labour-market-summer-and-september (released 2025-10-28). Both directly fetched and confirmed 2026-07-14 during adversary QA; supersedes this review's original WebSearch-only citation.

NEW-2026-5 — Entry-level hiring shift and active Canadian research initiative.

"Three in ten HR leaders in the United States and Canada report that their talent acquisition strategy is shifting toward hiring fewer entry-level workers in favour of mid-level employees." Also: "Signal49 Research, in collaboration with the Future Skills Centre, is examining if AI is already reshaping labour market entry in Canada, focusing on demand for entry-level jobs and the work requirements that shape their exposure to AI."

Source: WebSearch synthesis citing Signal49 Research, "AI and Entry-Level Labour Demand—June 2026," https://www.signal49.ca/in-fact/ai-and-entry-level-labour-demand_june2026/, and related HCAMag Canada reporting. Accessed via WebSearch 2026-07-14; not independently fetched from the Signal49 primary source within this review's budget — flagged ⚠️ still being checked, treat the "three in ten" figure as sourced-but-not-yet-independently-verified.

NEW-2026-6 — Ontario Basic Income Pilot cancellation and federal Bill S-233 death. Adversary QA pass, 2026-07-14: Wikipedia directly fetched and cross-checked; LEGISinfo directly fetched and confirmed.

"The Ontario Basic Income Pilot Project was a pilot project to provide basic income to 4,000 people in Ontario, Canada. In 2017, Ontario launched a three-year pilot enrolling roughly 4,000 people between 18 and 64 who earned below $34,000 individually or $48,000 as a couple... However, the project was terminated early by a newly elected Progressive Conservative government, and the final payments were made to participants in March 2019." Wikipedia's own text states the benefit in annual terms, not monthly: "Single participants received up to $16,989 a year while couples received up to $24,027... Those with disabilities would also receive up to $500 per month on top." $16,989/year ÷ 12 = $1,415.75/month (rounds to the commonly cited "$1,415/month"); $1,415.75 + $500/month disability top-up = $1,915.75/month (rounds to "$1,915/month") — the monthly figures cited in this page’s documents are arithmetically consistent with Wikipedia's own annual figures, confirmed by this review's direct calculation, though Wikipedia itself does not state the monthly figure directly and no government primary source for the exact monthly figure was located in this review. Separately, on the federal bill: LEGISinfo (Parliament of Canada, directly fetched and confirmed 2026-07-14) shows Bill S-233, "An Act to develop a national framework for a guaranteed livable basic income," sponsored by Sen. Kim Pate, reached "Consideration in committee" (Standing Senate Committee on National Finance) as its last recorded stage, spanning the 44th Parliament, 1st session (November 22, 2021 to January 6, 2025) — confirming the bill did not pass and the 44th Parliament's 1st session ended January 6, 2025 — consistent with this page’s "died at prorogation" framing. LEGISinfo's own page does not use the word "prorogation" (it states only that the session ran to January 6, 2025); the prorogation itself is independently well-established as a matter of public record (the Governor General prorogued Parliament January 6, 2025, at the sitting Prime Minister's request, extensively contemporaneously reported by Canadian outlets), so this is treated as a minor sourcing-precision gap rather than a substantive doubt about the underlying fact.

Source: Wikipedia, "Ontario Basic Income Pilot Project," https://en.wikipedia.org/wiki/Ontario_Basic_Income_Pilot_Project (directly fetched 2026-07-14; used here as an index to primary reporting, consistent with this project's practice of treating Wikipedia as a pointer rather than a primary source) and LEGISinfo, "S-233 (44-1)," https://www.parl.ca/legisinfo/en/bill/44-1/s-233 (directly fetched 2026-07-14). LegalClarity, "Canada Universal Basic Income: Bills, Eligibility and Costs," https://legalclarity.org/canada-universal-basic-income-bills-eligibility-and-costs/, remains WebSearch-sourced and not independently fetched in this review — its specific characterization of the bill's death as a "prorogation" event is flagged ⚠️ still being checked pending a primary parliamentary-procedure source (e.g., a Canada Gazette prorogation proclamation) confirming the mechanism by name.