The Hydrogen Center for Artificial Intelligence

NewsletterNo. 6

A new Congress, and what the layoff numbers really say

Control of Congress is changing hands in January. Here is what to watch on AI, and why the headline about AI layoffs is more complicated than it sounds.

By Yonathan Shanko, Founding Editor

The story: Congress changes hands

After the November 3 midterm elections, NBC News projects that Democrats will control both the House and the Senate when the new Congress begins in January. Euronews reported the same outcome. Before the election, Republicans held the Senate 53 to 47 and the House 219 to 214.

We do not take sides on elections, and we will not guess at what the new majority will pass. But a change in control does change who sets the agenda. The majority picks committee chairs, decides which bills get hearings, and controls which ones reach a vote.

What to watch on AI. A few threads from this fall are worth following into January:

  • Data center costs. Senate Democrats blocked the Ratepayer Protection Act in September, calling it too weak because it leaves the rule optional for states. Watch whether a mandatory version appears.
  • The White House accord. The voluntary AI safety accord signed in September leaves room to turn its steps into law. A new Congress may test whether anyone wants to.
  • Divided government. The White House has said it opposes new restrictions on AI development. Any AI bill will still need the President's signature or enough votes to override a veto.

Economy: AI layoffs, read carefully

AI has been the most-cited reason for US job cuts for much of this year, according to the outplacement firm Challenger, Gray & Christmas. That is true. It is also only part of the picture.

Challenger's September report shows employers announced 573,195 job cuts through September. That is down 39% from the same period in 2025. Employers cited AI in 120,136 of those cuts, about 21% of the total. Hiring plans for the year were up slightly.

So layoffs overall are down sharply, while the share blamed on AI is up. In its July report, the firm's Andy Challenger summed it up: "AI is shifting the labor market, it is not dismantling it."

The clearest damage so far is at the entry level. A Stanford Digital Economy Lab study found that workers ages 22 to 25 in the occupations most exposed to AI saw a 13% relative decline in employment, while more experienced workers in the same jobs held steady or grew. The declines were concentrated in jobs where AI does the work rather than helps with it.

One fundamental: "Cited" is not the same as "caused"

Layoff trackers like Challenger record the reason a company gives when it announces cuts. They do not verify it.

That matters because a reason is also a message. Telling investors that cuts come from AI efficiency can sound better than saying the company over-hired or demand fell. A company may also cut jobs to free up money for AI spending, even if no AI system is doing those jobs yet.

None of this means the AI numbers are fake. It means they measure what companies say, not what AI did. The more reliable evidence comes from studies like Stanford's, which track actual employment across millions of workers and compare jobs that AI can do with jobs it cannot.

Sources

If you have an evidence-based argument about AI, especially one that disagrees with ours, send us a pitch. Everything we publish is free. If you want to help keep it that way, become a supporter.