Welcome to this edition of JOBS THAT WORK, a policy and business newsletter about how everything is workforce, and how we can make everything better.

Employers are likely to use AI to justify shedding jobs that have been key pathways for workers to climb the economic ladder to better and more stable lives. That has big costs for employers as much as it does workers, and it requires different thinking in how policymakers design workforce programs and respond to AI-blamed job loss.

Something that’s always driven me nuts about some of America’s publicly funded workforce programs is that the law and underfunding drives them to aim too low in terms of outcomes.

Based on who qualifies for help under the law, federal workforce programs tend to serve the people American education systems and other pieces of society have strongly suggested Go Do Something Else without offering a clear way to the labor market. Putting whatever shine is needed on that person to get a job that helps them get ahead of the cost of living can take money and time. Congress strongly favors moving bodies and paper as quickly and cheaply as possible through federal workforce programs, which tends to foreclose effective long-term solutions.

In other words, it costs too much to get real results. Where workers who complete many of these programs tend to land is in the general vicinity of a job that might help them get ahead—one day. Very often public workforce programs deliver their graduates to the type of jobs that are entry-level for people who had more resources growing up or just took to typical schooling better. Or the programs deposit them in a job a couple rungs below those jobs because, well, it’s available right now.

Most times, providers don’t have much of an idea of how that worker will get from the place they leave them to being able to comfortably afford an increasingly expensive life. Some workforce providers and even some policymakers insist it’s the best they can do with the tools they have. Some workers need stability first, they say. Better-paying employers, they say, need proof that a person will “work out” for reasons that my lawyer brain can find hard to separate from the worker’s class or race or sex.

The problem is that those jobs really can’t keep being the “best” place to leave those workers. Not anymore. As documented quite well in research published last week by Opportunity@Work and Brookings, many of the kind of jobs public workforce programs aim their workers at are starting to go away.

And they’re also going away for the people who have been “safe” hires for them for generations. Entry-level job postings have declined 35 percent since 2023, and new college graduates are having a tough time getting hired. Employers are blaming AI, even if the jobs aren’t eventually filled by AI.

In other words, not only are we losing the typical weigh stations we place workers who need more help getting into jobs that help them get ahead, we’re likely going to have an entirely new class of people who can’t reach the ladder to get into jobs that pay well and treat them well enough to build a stable life. That has a cost to employers as well as workers, and we need to rethink the shape and aim of publicly funded workforce programs, as well as how involved policymakers and elected leaders are willing to become on AI-related job loss.

O@W has been at the forefront of work to remove things like unnecessary college degree requirements that can keep more-than-qualified workers from getting promotions or filling jobs that help them get ahead. O@W calls these workers “STARs,” or Skilled Through Alternative Routes. Often STARs pick up or refine these skills in what O@W calls “Gateway” jobs, or roles where they can build skills that let them translate into jobs that can help them get ahead.

Per O@W and Brookings research, 15.6 million STARs are “highly exposed” to AI-related job loss, and almost half of the jobs between those “Gateway” gigs and higher-paying “Destination” roles are as well.

"It's just a much broader set of occupations and workers that are going to be impacted," said Justin Heck, a co-author of the report and O@W’s senior director of research data, when we chatted last week.

Heck has concerns not just about the doorways closing for populations to get ahead, but also for employers to find the talent they need and future leaders of companies to pick up the skills needed to run their businesses. A lot of skills vital to doing good, effective work attach to workers as they climb up the ladder. Companies, though, are looking to AI for more “frictionless” processes and closing off pathways by which people learn to do a good job in those "Destination” roles.

On my end, I’ve been hearing about decisionmakers at employers that seem increasingly detached from how they actually find people and overly ambitious as to how AI can just let them gloss over recruiting and training talent. Many are facing shortsighted expectations from investors that they wipe out as much staff as possible due to AI to goose profits. Based on my conversation with Heck, I’m also worried that shoot-before-you-aim mass AI layoffs might not just harm workers, but the long-term prospects of companies laying them off due to the diminished seasoning of their future leaders.

If employers are wiping out entire taxonomies of workers in hopes AI can pick up the slack, those workers will need to go somewhere to keep getting ahead—if only to keep our consumption-driven economy afloat. Moving toward more skills-first hiring strategies is a good approach "instead of [routing AI-affected workers to] ground zero and not assuming they have valuable experience,” Heck said.

That said, research suggests employers who have made efforts to hire based on skills, not degrees or other pedigrees, haven’t exactly gotten the message through to their hiring managers.

To be clear if you haven’t noticed, this isn’t a problem with AI, the technology. This is a problem with people who make decisions on it.

"I know that employers might want to make us feel like AI adoption is gravity, but these are collective choices that we make,” Heck said.

America’s top workforce policymakers are responding to this layered, possible existential problem by… not doing anything that could really address it. No one expects anything meaningful from Congress on these topics. The Trump Administration has done some interesting Generalized AI Workforce Stuff in recent weeks. The most creative thing it’s done in this space mostly tries to make AI friendlier for workers, and this year it’s tried to minimize concerns about AI-related job loss by saying it is “fearmongering” and “theoretical.” Staying on that path risks being dismissed for being tone-deaf and out of touch with a serious problem. I think that effort is well-intended in fact but hamstrung by the policy limitations of the current administration.²

That likely leaves it to state and local leaders to fill the gap, who unfortunately already have too full of a plate. What should they add to it? Well, public workforce programs really shouldn’t just deposit a worker at the nearest job and hope everything works out. They have to think in terms of ladders and options—how a job can lead to another job for the worker that helps them stay ahead, and what ladder that worker can jump to if technological change tips the first (or second) one over.

Separately, governors and mayors need to be proactive in working with employers to not just fire everybody because they think AI lets them. That has a cost to the employers down the line and state and local economies, and proactive negotiations and nudges (like tax benefits) are needed to ensure that this transition is done more thoughtfully than urged by investors hooked on Magic Beans.

Politically, I think that’s the best they can do for now, particularly since neither major party’s national leadership seems like they might meaningfully intervene at the moment. That said, to the extent allowable, the potential for calamity seems dire enough—and the decisionmaking underlying it thin enough—that it may be nearing the time to have a conversation about the word that starts with “r” and ends with “-egulation” of AI-related layoffs.

That would be probably be politically popular but hard policy to make, and frankly, it may not end up in a great place after litigation. But the risks here are too great not to try something other than waiting for the worst to happen.

So hey, the White House released a budget calling for the end of everything… again. I’ll have some more thoughts on that on Thursday in THE MONEY because I’ve seen some analysis that’s a bit… ambitious in terms of how this is likely to be received on the Hill, where the White House budget office isn’t really well-liked at the moment. Also, this week, I’ll be adding some new programs—and warnings—over at The Federal Funding Library.

In case that wasn’t disruptive enough, we could be closing in on the President firing Labor Secretary Lori Chavez-DeRemer. Suspect I’ll have thoughts on the impacts of that, too.