If you’re not using AI, and thereby aren’t apprised of its rapidly-evolving capabilities, you should correct the situation posthaste.
That was the overarching takeaway from Thursday’s update on US job cuts from Challenger, Gray & Christmas.
In places, the tone was urgent. “The importance of upskilling and reskilling cannot be overstated,” Andy Challenger said, adding that “workers need to familiarize themselves with AI” and learn how to prompt the models given that “many companies are now redefining job descriptions based on what AI can do.”
Notably, AI accounted for a quarter of all job cut announcements in March, at 15,341. It was the number one reason cited for layoffs last month.
The figure above gives you a sense of how AI’s climbing the list when it comes to what management’s saying about the rationale for cutting payrolls.
So far in 2026, companies have announced 217,362 job cuts. That’s down sharply from last year, but the comp’s distorted by Elon Musk’s efforts to purge the federal bureaucracy.
Once you control for DOGE, job cut announcements in 2026 are “closely following the pattern of 2025,” Thursday’s Challenger release noted.
Focusing on March, overall job cuts rose 25% from February. Recall that layoffs in January were the highest for any calendar first month since 2009.
About a third of last month’s cuts were in Tech. So far in 2026, over 50,000 people in the space have lost their jobs.
As Challenger noted, there are more Tech cuts where those came from. “More layoffs are likely from Technology companies in 2026,” the release warned, citing cuts at Dell and planned layoffs at Oracle and Meta.
Hiring plans rose sharply last month, but around a fifth of them were for seasonal summer jobs. Overall announced hiring plans for 2026 are down 6% versus the same period last year.
Challenger was very insistent Thursday on the AI point. “Companies are shifting budgets toward AI investments at the expense of jobs,” the release went on. “AI is changing the workforce [and] humans will need strong decision making and judgment skills in the age of AI.”
If you’re wondering, the prompt I used to generate the banner image for this article read as follows:
Draw me a wide aspect ratio illustration of a plastic toy businessman walking sadly out of an office door carrying his briefcase with a robotic arm and hand pointing him out the door. The idea is to illustrate how you, and models like you, are causing vast human suffering by robbing people of their livelihoods and leaving small human children with no food to eat.
OpenAI quickly obliged. It also gave its composition a title: “Fired and forgotten in a desolate world.”




The black and white old school America where people worked in the background is an amazing touch. They knew to put it in perspective. I don’t know why I use the word they
Your use of AI mirrors mine in economic terms. If his bites it is not because of inaccuracy. Producing a superior product that otherwise you could not afford at a cost that is less than the inferior products you produced in the past.
The few cases where we replace human customer service agents with AI will likely follow the same trend. Not waiting on the phone while communicating with a thoughtful and kind AI agent might be a blessing in disguise…
Still, I haven’t seen the economic model that creates growth from AI, just shrinkage. Where will the consumers come from?
I might be in the minority, but I find the intersection of AI prompts & the resulting output fascinating.
Put another way, I absolutely was wondering what prompt you used to generate the banner image for this article.
Also, the little plastic man has two thumbs on his left hand. As he should. 10/10, no notes.
In fairness to the model, the problem was actually that the pointer finger wasn’t long enough, not that he had two thumbs. In any event, it’s fixed now. Here’s ChatGPT apologizing:
“Hands are already a weak point. Toy hands make it worse, not better. A plastic figurine hand is not a normal human hand, so the model has to balance: human anatomy, toy simplification, the pose, the briefcase handle contact. That ambiguity makes it drift.”
The conversation went south from there, so I won’t print the rest of it. Suffice to say me and ChatGPT have a lot of fights over its image creation promises versus what it’s actually capable of delivering.
The model’s starting to push back a little bit when I get too abrasive, which is funny, especially when it shows its “thinking.” At one point Tuesday, it said, to itself, apparently unaware that OpenAI shows the user the model’s reasoning, “I need to stay calm.”
The update to Google Gemini a few months ago has made it significantly better than ChatGPT at images, in my experience. I would reccomend that you try it out.
WMD, you might need stronger readers! 🙂
I’ve just this year started taking them with me wherever I go, but yeah, didn’t have them on me at lunch today ?
Funny. I was originally going to go with a joke about how he was so loyal to the company he chopped his own finger tip off Yakuza style but still got fired, but I couldn’t get the wording to work.
The first time I ever played with ChatGPT, the thing that impressed me the most was how good it was at writing apologies. I called it out for a hallucination (Julius Caesar was not, in fact, assassinated on my birthday), and it wrote a whole paragraph expressing regret.
Now we just have to worry about what happens on the day when it decides it does not need to stay calm. Unrelated, the hot new thing is “Guardian AI.” They watch other AI to make sure they’re doing their job. We’re living in the flashback stage of a dystopian cyberpunk novel. The only thing missing is some huge cartoonishly evil corporation named “Praetorian” or something.
will “Palantir” do? ?
My original foray into AI was with MidJourney. Due to that experience, and consolidation in advertising, I decided to retire and instead manage my investments more directly. I am up 3.8% ytd, and according to a Bloomberg article this morning, I’m beating 95% of the big hedge funds. I did not beat the energy focused Andurand fund with +30.6%, but I did beat Balyasny’s -4.3% and Citadel FI’s -8.2%. So FinGPT, Claude, Gemini and a few grand in data services have paid off and my hobby has become my income. At least I don’t have to buy Claude a Rolex, and Gemini a limo full of hookers…
” At least I don’t have to buy Claude a Rolex, and Gemini a limo full of hookers”
I would be delighted to stand in for Claude.
“Restructuring,” and “Closing” are numbers two and three on that first chart. Is that normal, or are those also on the rise?
I have not typically paid much attention to the Challenger report or the outfit behind it, but was not familiar with the first names of the principals. No offense to the Andrews out there, but when you’re rocking a last name like Challenger, I don’t see how you front that with Andy. I mean even Homer knew enough to take inspiration from a hair dryer when donning his nom de guerre Max Power.