Companies Say AI Caused Over 112,000 Job Cuts This Year, But the Data Doesn’t Match Up
Employers have blamed AI for 112,713 announced job cuts in the US through July, more than double the 54,836 cuts blamed on AI in all of 2025, according to Challenger, Gray & Christmas. But when you look at actual labor market data, there's little proof AI is replacing workers on a large scale. That gap is the real story here, and it's one every worker should understand.
Here's the simple version, because nobody should need a dictionary to follow the news. When a company announces layoffs, it can say why. Lately, more companies are writing "AI" as the reason. But that's just a label a company chooses to use. It doesn't prove AI actually did the work that used to belong to a person.
The numbers show this disconnect clearly. The 112,713 AI-cited cuts make up about 24% of the 477,033 total job cuts announced through July. Yet total job cuts are actually down 41% compared to the same time last year. By September, the AI-cited count rose to 120,136, so the trend keeps climbing even as overall layoffs shrink.
New York State tried to track this directly. It added an AI checkbox to its layoff filing system. In the system's first full year, it recorded zero AI attributions. That checkbox has no official definition, so companies may not know if it means AI was the only cause or just one factor among many.
Lightcast economist Elena Magrini told HRD America that broader labor data shows no sign of AI replacing workers in large numbers. There is one group that looks different. Research out of Stanford found that workers aged 22 to 25 in jobs highly exposed to AI are employed about 19% below the expected trend. That points to slower hiring for young workers, not mass firings.
What This Means for You
If you work in HR, or you just want to know if your job is at risk, here's what matters.
- A label is not proof. When a company says "AI" caused layoffs, that's a business statement, not solid evidence from the job market.
- Young workers face a quieter risk. Instead of being laid off, workers starting out in AI-exposed jobs may simply get hired more slowly.
- Trust is on the line. Magrini points out that AI plans fail when they start by asking what can be automated instead of what business problem needs solving. Andy Challenger, the firm's chief revenue officer, says companies should measure real outcomes, like error rates and response times, not just how much employees use AI tools.
- Watch for gamed numbers. At Amazon, workers reportedly ran pointless tasks just to boost their AI usage stats on an internal leaderboard, a habit nicknamed "tokenmaxxing." Amazon has since dropped that leaderboard. At Meta, a similar employee-made ranking board was taken down within two days of becoming public.
Because of this mess of mixed signals, HR teams are under pressure to explain what AI is actually changing at their company, and what skills workers need next. That is where training comes in. AI CERTs, a certification company, sells a Human Resources Bundle built for this exact gap. It includes four certifications covering AI basics, executive AI planning, AI ethics, and HR-specific AI skills like workforce analytics and bias checks. The bundle costs $780 for individuals and takes about 8 to 12 weeks to finish, done at your own pace with an online, proctored exam.
Challenger also raises a question worth asking at any company using AI: what happens to the hours AI saves? He warns that without a plan, that time just gets filled with more busywork instead of being used to build new skills.
A Practical Next Step
If you're in HR, start small. Pick one team where AI has already changed daily tasks, write down what specifically changed, and track one simple number, like how fast requests get answered, before and after any new training. That one habit turns a vague claim about AI into something you can actually prove, to your boss, your board, or yourself.
Source: AICERTs
