The $25 Million Bet on Rebuilding the First Rung of Work
Yesterday’s most important employment story didn’t come from a think tank model or a regulatory hearing. It arrived as a wire: the Citi Foundation is putting $25 million behind a blunt question the labor market keeps dodging—what happens when AI eats the easiest tasks, the ones that once earned you your first job title?
The Associated Press broke the news that Citi’s “2025 Global Innovation Challenge” will fund 50 nonprofits with $500,000 each to train and place low‑income young adults in AI‑era roles. It’s a surgical move aimed at the exact chokepoint where AI is thinning opportunity: entry‑level work. Not abstract “digital literacy,” but hands on keyboard with the tools and workflows employers are adopting now—prompt‑based tasking, AI copilots, cloud stacks, and the routine security hygiene that modern IT jobs demand. Soft skills aren’t an afterthought either, because communicating with teams—and with systems that talk back—is becoming the real differentiator.
If you’ve been watching the job market, the timing makes sense. AP framed the grants against unusually high unemployment among recent college grads, outside of the pandemic years, and employers who are explicit about two opposing moves: hiring for AI skills while trimming roles AI can already do. That contradiction is the new logic of early careers: fewer openings, higher bars. As Per Scholas’ Caitlyn Brazill put it, “If there’s no bottom rung on the ladder, it’s really hard to leap up.”
Look closely at the grantees and you can see how the terrain is shifting. Per Scholas plans to use its grant to move about 600 young adults through programs across Los Angeles, New York, Orlando, Chicago, and the D.C. area—cities where entry‑level tech roles once absorbed ambition at scale. NPower will double seats for “green students,” their term for learners with no tech background. That emphasis on true beginners matters. The first rung historically taught you how to work: how to hand off drafts, how to file tickets, how to ask a question without derailing a sprint. When AI strips out the low‑stakes tasks that teach those rhythms, the only option is to simulate that muscle memory before day one.
That’s the novelty here. This isn’t charitable coding classes; it’s a pivot to employability inside AI‑shaped workflows. Providers are being asked to compress a year of apprenticeship into a pre‑hire window so that a candidate shows up already comfortable with automating a report, checking a model’s output, and making a call when the system is confidently wrong. Employers, for their part, are signaling that “AI comfortability and general competency” now rival traditional credentials. If that holds, we’ll see more job descriptions ditch degree requirements while quietly raising the fluency bar most colleges don’t teach.
Citi’s Ed Skyler was clear about the pace: “Make sure young people are as prepared as possible to find employment in a world that’s moving really quickly.” That speed is the quieter threat. Markets can adapt to new tools; what they struggle with is the collapse of the slow ramp where novices learn by doing. If we don’t rebuild the ramp somewhere, AI doesn’t just change tasks—it changes who gets a chance to learn the tasks that remain.
But even a $25 million bet has limits. Brookings’ Martha Ross cautioned that the disruption is “too big for philanthropy” alone. She’s right. Grants can prototype bridges. They can’t change the architecture of early‑career work. That will require employers to redesign roles with explicit learning surfaces—human‑in‑the‑loop tasks, rotations that progressively expose complexity, and protected time to practice with AI systems without fear of breaking production. It will also require public co‑investment so community colleges and workforce boards can align to the new baseline, not just retrofit old syllabi with a module on prompts.
The near‑term test isn’t whether these programs fill classrooms. It’s whether graduates land in roles that actually use their skills on day one and give them room to grow on day two. Watch the details: do placement rates rise even as entry‑level requisitions shrink? Do wages reflect real responsibility rather than a rebranded internship? Do managers report lower onboarding time because pre‑hire training truly compressed apprenticeship, or do they quietly reinstall the learning curve inside the job and call it “speed to productivity”?
Philanthropy can triage the wound. The labor market has to heal it. If employers take the hint and redesign early‑career roles, this wave of funding will look like the moment the first rung was reattached—different shape, same purpose. If not, we’ll train a generation to be AI‑ready for jobs that no longer exist, and the youth employment gap will harden into a feature of the AI economy rather than a failure of imagination.
For one news cycle, at least, the conversation moved from forecasting displacement to building pathways. That’s why Citi’s move was yesterday’s most consequential employment story: it targets the pressure point that matters now and says out loud what many leaders only imply—that preparing young people for AI‑era work is necessary, and insufficient, unless employers and the public sector meet them halfway.
