Library: Case Entry
RAISE US has committed more than $500 million to retraining American workers for an AI-driven economy. Retraining only works if there are durable jobs waiting on the other side. Algorism asked RAISE US which occupations displaced workers are expected to move into. It did not respond. This is the record.
RAISE US is a workforce coalition launched on 25 June 2026, backed by major AI developers, technology companies, employers, philanthropies and bipartisan state governments. It launched with more than $500 million committed toward a stated $1 billion goal, with initial state partnerships in Arkansas, Connecticut, Maryland and Utah.
Its public materials set out the problem in its own terms. An estimated 50 million American workers hold jobs vulnerable to AI. Studies it cites estimate that between 7 and 11 million American workers could be displaced. It describes the training systems, funding structures and career transition infrastructure those workers would need as not built for this moment, and says it is building them.
Its launch release states that the companies deploying AI have the most direct line of sight into where jobs are changing. Its strategy centres retraining, redeployment and worker transition, and it states that success will be measured by whether workers obtain and keep good jobs. Its homepage carries a partner statement describing the question the coalition convenes around: "where are the AI-resilient jobs, and how do we move people into them quickly."
The coalition also included analytical capability from the start. Its launch release names the Burning Glass Institute among its initial partners and its president, Matt Sigelman, on its advisory board, and describes that institute's contribution as labour market analytics and AI impact analytics. Algorism records that as a fact about the coalition's composition at launch and draws no inference from it about work completed.
On 17 August 2026, Algorism captured the coalition's public materials, including the homepage, About, Policy Lab and Press pages and the launch release, as a dated evidence baseline. Each captured file carries a SHA-256 hash recorded at capture time and retained by Algorism, so any later version can be compared against what was published on that date.
Retraining is a bridge. A bridge needs a far bank. As of the 17 August 2026 capture, the coalition's public materials did not identify the occupations displaced workers are expected to be retrained into. They name sectors and mechanisms: healthcare, education, advanced manufacturing, apprenticeships, short-term credentials, career navigation, entrepreneurship, wage support and internal redeployment. Sectors are not occupations. A hospital employs surgeons, nurses, coders, billing clerks, receptionists, technicians and administrators, and those jobs have radically different exposure to automation.
Before committing hundreds of millions of dollars to retraining, there presumably exists analysis showing that enough durable jobs are expected to exist on the other side. The question is whether that analysis exists and whether the public may see it. Specifically:
Institutions that ask for public trust on the strength of what they can see owe the public a look at what they saw.
Training people for work does not create the work. Without evidence that enough durable destination jobs exist, large-scale promises of retraining risk functioning as damage control: reassurance offered to workers facing displacement before anyone has shown where the replacement jobs will come from. That is the concern this entry tests, and the reason the six questions above are the ones that were asked.
The problem is not new. When coal employment collapsed across parts of Appalachia, coding was widely promoted as a path for displaced miners. One heavily publicised programme, Mined Minds, drew national coverage and received a share of a $1.5 million Appalachian Regional Commission grant. A 2019 New York Times investigation reported that almost none of the enrolled miners obtained programming jobs. Former students later sued, alleging the programme's promises amounted to fraud, and their attorney told a local newspaper he had found no graduate placed in the technology industry apart from two brought back as trainers. The organisation's co-founder disputed the characterisation and attributed the outcomes partly to students' expectations, and the allegations were not proven in court.
The lesson of that episode is not that retraining is useless, and it is not that RAISE US is Mined Minds. The two are not comparable in scale, governance or resources, and nothing here suggests otherwise. The lesson is narrower and it applies to any transition plan: some of the miners did learn to code, and the local jobs still were not there. A credible transition plan needs both sides of the bridge. RAISE US has shown the scale of the retraining commitment, but it has not publicly shown the destination jobs at a scale that matches that commitment.
On 17 August 2026, Algorism captured and hashed the coalition's public materials as an evidence baseline.
On 18 August 2026, Algorism sent a pre-publication inquiry to the coalition's published media contact, copying its general contact address. The inquiry asked the six questions above, requested a link or copy of any occupational forecast, workforce model or supporting research, and asked, if that analysis was not complete, how large-scale retraining was determined to be the appropriate response before destination occupations were identified. It stated that publication was expected in September, that any response would be represented accurately, and that if research had been overlooked or the approach misunderstood, Algorism wanted to know before writing.
On 28 August 2026, a follow-up was sent within the same thread, restating the question and setting a response date of 7 September 2026. It stated that if no response was received, the published piece would note the dates of both inquiries and that none was received. Both messages are reproduced verbatim in the appendix below.
So that readers can verify the goalposts never moved, the readings below were written on 18 August 2026, the day the inquiry was sent, before any response could arrive. They are reproduced here verbatim.
A response that engages the destination-occupation question substantively, even with uncertainty, even in part, is a good outcome, and this entry would say so. Algorism applies its own rule here: an inquiry capable only of recording failure is hunting, not documenting.
As of 16 September 2026, following the 18 August inquiry and the 28 August follow-up, RAISE US has not responded. That is recorded as a dated fact and is not characterised further. Algorism does not know why no response was sent, and does not speculate. New organisations receive substantial correspondence, and an unanswered email is not evidence of anything beyond itself.
The inquiry remains open. A response received after publication will be added here with a dated changelog entry, in full or by link.
On the same date, Algorism searched for the analysis again. The search covered the coalition's public site, including its homepage, About, Government, Employer, Education, Policy Lab and Press pages, its launch release, announcements by coalition member organisations, and general web search. As of 16 September 2026, Algorism located no published RAISE US analysis identifying the destination occupations expected to absorb displaced workers, their capacity, expected wages, transition requirements, or expected durability as AI capabilities improve. The scope of that search is stated so that readers can reproduce or refute it.
One development between 17 August and 16 September 2026 bears on the question and is recorded here.
On 7 September 2026, Cognizant, a RAISE US member, published an AI-era workforce announcement stating that it plans to hire 1,500 US college graduates and scale two AI-era job categories, Frontier Certified Engineer and Frontier Business Operator, to a combined 15,000 people. Those commitments were made earlier and separately. Cognizant created the two job categories on 1 June 2026, announced scale targets of 5,000 engineers and 10,000 business operators on 9 July 2026, and announced the 1,500 US graduate-hiring target on 15 July 2026; its chief executive later told Newsweek that roughly 600 of those hires had already been made. The 7 September release brought those existing commitments together. The 15,000 figure describes the intended size of a job family rather than a count of new US positions, and the announcement does not state that all 15,000 are new US jobs or that they will be filled in 2026. Independent reporting drew the same distinction. This is concrete occupational detail from one participating employer. It is not an analysis of where displaced workers across the economy are expected to go.
Otherwise the coalition's public materials are substantially unchanged from the 17 August baseline. The displacement figures, the four focus areas, the statement that the organisation is at the beginning of its work, and the convening question about where the AI-resilient jobs are all remain as captured.
This entry does not claim the coalition's commitment is insincere. It does not claim retraining is the wrong response. It does not claim no analysis exists; it records that none was located publicly and none was provided on request. It does not claim to know why no response was sent. It does not assess, score, rank or rate RAISE US, its members, or any individual, and it will not become a recurring watch.
It documents three things: what was publicly committed, what was asked, and what was answered. Readers can weigh the rest themselves. The one position Algorism takes is the one already in its doctrine: institutions that ask for public trust on the strength of what they can see owe the public a look at what they saw.
| Claim | Source | Retrieved |
|---|---|---|
| Launched 25 June 2026; more than $500 million committed toward $1 billion; state pilots in Arkansas, Connecticut, Maryland, Utah | RAISE US launch release | 17 Aug 2026 |
| 50 million American workers hold jobs vulnerable to AI; 7 to 11 million displacement estimate attributed to cited studies | RAISE US homepage | 17 Aug and 16 Sep 2026 |
| Member companies have the most direct line of sight into where jobs are changing | RAISE US launch release | 17 Aug 2026 |
| Success measured by whether workers obtain and keep good jobs | RAISE US launch release | 17 Aug 2026 |
| Convening question about where the AI-resilient jobs are | RAISE US homepage, partner statement | 17 Aug and 16 Sep 2026 |
| Burning Glass Institute named an initial partner and Matt Sigelman an adviser at launch; role described as labour market and AI impact analytics | RAISE US launch release | 17 Aug and 16 Sep 2026 |
| Mined Minds received a share of a $1.5 million Appalachian Regional Commission grant; 2019 New York Times investigation reported almost no enrolled miners obtained programming jobs | New York Times, 12 May 2019, as reported in subsequent coverage; grant detail | 16 Sep 2026 |
| Former students' lawsuit alleging fraud; attorney's statement on placements; co-founder's response; allegations unproven in court | Belt Magazine; CBC News | 16 Sep 2026 |
| Inquiry sent 18 Aug 2026; follow-up 28 Aug 2026; response date 7 Sep 2026; no response as of 16 Sep 2026 | Algorism correspondence record, retained in full | 16 Sep 2026 |
| Cognizant: 1,500 US graduate hires and 15,000 combined Frontier job family, presented together 7 Sep 2026 | Cognizant release, 7 September 2026 | 16 Sep 2026 |
| Two job categories created 1 Jun 2026 | Cognizant release, 1 June 2026 | 16 Sep 2026 |
| Scale targets of 5,000 engineers and 10,000 business operators announced 9 Jul 2026 | Cognizant release, 9 July 2026 | 16 Sep 2026 |
| 1,500 US graduate-hiring target announced 15 Jul 2026; roughly 600 hires reported as already made | Cognizant release, 15 July 2026; Newsweek interview with Cognizant chief executive | 16 Sep 2026 |
| 15,000 is a job family size, not a count of new US openings | Independent reporting on the same announcement | 16 Sep 2026 |
The 17 August 2026 capture is retained by Algorism with a SHA-256 hash for each file, so any subsequent change to the pages cited here can be demonstrated rather than asserted. It is held as a working record rather than published as a downloadable package; it is available on request for anyone checking a disputed claim.
Anthropic, the developer of the Claude models, is a funder of RAISE US. Claude models participated in drafting this entry and the underlying inquiry, and that conflict was disclosed inside the working process from the first session. The analysis was cross-checked throughout by a model from a different developer, and the human editor held final judgement. The full working thread is retained.
The text of this entry is published under a Creative Commons Attribution 4.0 International licence and may be shared and adapted with attribution.
This entry is dated and not edited silently. When it changes in substance, the change is logged here.
v1.0, 17 September 2026. First publication. Evidence current as of 16 September 2026. Records the 17 August baseline, the 18 August inquiry, the 28 August follow-up, the preregistered readings, the absence of a response as of 16 September, and the Cognizant announcement of 7 September. Section 3 sets out why the destination question is the one asked; it states a general condition and makes no claim about RAISE US's intent.
18 August 2026. Subject: Request for information before publication: RAISE US occupational transition analysis.
Dear RAISE US Team,
My name is John Jerome. I am the founder of Algorism, a project of The Great Unplugging Inc., a 501(c)(3) nonprofit examining how people and institutions should respond to the social and economic changes advanced AI is creating.
I am preparing a piece on AI-driven workforce displacement that discusses RAISE US. Before publishing, I want to be sure I understand your strategy correctly and give RAISE US the opportunity to provide information I may have missed.
I have reviewed your public materials on the more than $500 million committed toward preparing American workers for AI-driven changes in employment. RAISE US has stated that participating companies have some of the most direct visibility into where jobs are changing, and has made retraining and worker transition central to its response. The one part of the strategy I have not been able to locate is the occupational destination analysis.
Specifically, has RAISE US developed an occupational transition model estimating: which occupations are likely to experience substantial AI-driven displacement or contraction; which occupations are expected to absorb those displaced workers, and approximately how many additional workers each can support; the training, time, and cost required for those transitions, and the wages workers could reasonably expect afterward; and why those destination occupations are expected to remain viable as AI capabilities continue to improve?
I understand that the exact effects of AI cannot be predicted and that pilot programs may be needed to determine which training approaches work best. My question is about the step before that: the analysis indicating that sufficient durable employment opportunities are likely to exist for displaced workers to move into.
If RAISE US has published that analysis, or the supporting research behind it, I would appreciate a link or copy. If it has not yet been completed, I would appreciate understanding the reasoning that led RAISE US to make retraining a central strategy before the destination occupations and their capacity were established.
I am approaching this as an evidence question rather than starting from a conclusion. If I have overlooked relevant research or misunderstood your approach, I want to know that before writing about it. I expect to publish in September and will represent any response accurately; if none is received, the piece will note that as well.
Thank you for your time.
Best regards,
John Jerome
Founder, Algorism
A project of The Great Unplugging Inc., a 501(c)(3) nonprofit
28 August 2026. Sent within the same thread.
Dear RAISE US Team,
I am following up on my email of 18 August, which asked whether RAISE US has developed an occupational destination analysis behind its retraining strategy: the specific occupations displaced workers are being trained to enter, how many workers those occupations can absorb, expected wages and training requirements, and why those occupations are expected to remain viable as AI capabilities improve.
I have not yet received a response. I expect to publish in September and want RAISE US's position represented accurately. Anything you can provide by 7 September would be welcome: the analysis itself, a pointer to where it is published, or an explanation of the strategy's basis if the analysis is still in development.
If no response is received, the piece will simply note the dates of both inquiries and that none was received.
Thank you for your time.
Best regards,
John Jerome
Founder, Algorism
A project of The Great Unplugging Inc., a 501(c)(3) nonprofit
Both messages were sent to the coalition's published media contact address, copying its general contact address, from an algorism.org address.