On August 19, New York’s governor’s office announced that its FutureWorks Commission had begun holding listening sessions. The release runs a few hundred words. One sentence in it is load-bearing: in New York, women make up 84% of employment in administrative back-office roles.
The first session seated women doing exactly that work. Administrative, clerical, customer service. The release names all three, and states plainly that some experts expect AI to shift or absorb tasks in those roles, roles that have historically been a pathway to the middle class.
That framing is rare in American state AI documents. Most of them are about model safety, data governance, and compute investment. This one names an occupation, then names who holds it.
What the commission is
Twenty members. Three co-chairs: former U.S. Labor Secretary Tom Perez, TIAA’s Thasunda Brown Duckett, and Breakwater Initiative founder Molly Kinder. The membership spans technology, workforce development, education, and economics.
It owes two deliverables. A method for monitoring AI’s employment effects on real-time data, and a set of recommendations for both government and the private sector. Both are due by the end of 2026.
More sessions follow over the coming weeks, widening to workers of other backgrounds, small businesses, and other stakeholders.
The same release lays out the other half of the ledger. Empire AI is a $500 million partnership of public and private universities building an academic computing center at the University at Buffalo. The state has extended AI training and tools to its own workforce of more than 100,000 employees. Its Office of Information Technology Services wrote the first statewide policy in the country on acceptable AI use, then followed it with governance guidance for agencies.
New York is not asking whether to adopt AI. It answered that, at scale, and the commission exists to handle what the answer leaves behind.
Three clocks, running at different speeds
Line the timetables up and the gap is the story.
The corporate clock runs fastest. On May 19 last year, at its Hong Kong investor day, Standard Chartered attached 7,000 positions, 15%, and 2030 to the same sentence, cutting into a 52,271-person corporate-function workforce. A bank said out loud what it intends to do to its back office, and booked it through the end of the decade.
The legislative clock runs second. On August 14, Korea’s National Assembly took delivery of a three-bill package that charges the employer whose AI adoption cut the headcount, not the vendor who sold the model. Whether it passes is a separate question. It is drafted text with a collection path attached.
New York’s clock runs last. Sessions start in August, recommendations land in December, and any statute waits for the 2027 legislative session. That is more than a year from today.
The back-office workload will not hold still for a year while the paperwork catches up.
Why 84% is the number to keep
It converts something vague into something checkable.
The common shorthand for AI exposure is “low-skill jobs” or “entry-level jobs.” Both are wrong in a way that matters. On August 12, Stanford’s Digital Economy Lab widened its young-worker gap to 19% and named the actual mechanism: what determines exposure is not skill level but how much of the work is codified, written down, and transferable. Administrative back-office work sits at the top of that distribution among white-collar occupations.
Korea’s July microdata pointed the same way from a different direction. The center of gravity of youth job losses moved out of manufacturing and hospitality and into information & communications and professional services, which is to say out of blue collar and into knowledge work.
New York adds the third axis: who is standing in the exposed occupation. At 84%, a multi-year contraction in administrative headcount does not distribute evenly across the labor force. It concentrates. That observation used to live in academic papers. It now lives in a governor’s press release, which is a different kind of document with a different kind of afterlife.
Public sentiment has already moved. Pew Research Center reported on August 18 that 71% of U.S. adults expect AI to reduce the number of jobs over the next two decades, up from 64% in 2024. Among adults under 30 the figure moved from 61% to 73%.
Who this release is actually for
People currently in administrative, clerical, and customer service roles. The three words the governor’s office chose are the coordinate system for this cycle. A quick way to locate yourself in it: could your daily output be handed, complete, to someone who has never met your colleagues and never logged into your systems? If yes, you sit on the codified side. If no, you are holding tacit context, which is the only thing currently holding.
The state’s own 100,000 employees. New York State is one of the largest administrative employers in New York State. It has given itself the training, the tools, and the usage policy. What happens to that headcount over the next two years is the cleanest leading indicator available anywhere in the state, because it is not driven by quarterly earnings and its numbers are public.
Anyone planning to enter the workforce through an administrative role. The release uses the past tense: roles that “have historically been a pathway to the middle class.” Harvard’s tracking of 62 million workers caught the same shape earlier, that AI is not firing juniors so much as declining to hire them. Back-office work is where that logic bites hardest, because low entry barriers and high codifiability arrive in the same job.
Anyone waiting on policy. December for recommendations, 2027 at the earliest for law. The window in between belongs to no document.
State Senator Jessica Ramos supplied the honest line in the release: the measure of New York’s success with AI cannot only be how fast it adopts the technology. She is right. On the evidence of the three clocks, speed is currently a one-sided contest.
Sources:
- Office of Governor Kathy Hochul: Governor Hochul Launches FutureWorks Commission Listening Sessions as Part of Plan to Address AI Workforce Impacts Across New York (August 19, 2026)
- Pew Research Center: Young adults in the U.S. are increasingly wary of AI, concerned it will take jobs (August 18, 2026)
- Axios: Hochul launches AI workforce push with focus on women (August 19, 2026)
