One in 11 Jobs? What the AI Layoff Numbers Really Say

If AI news has been tightening your stomach lately, you are in the right place. One week the headline says millions of jobs are going away. The next week, companies are calling their old workers back. It is a lot, and most of it never tells you what to do next. So today, we are going to slow down and look at what the numbers really say.

This is AI Fear Check, episode one. This show is researched with AI and uses a generated AI voice of Shannon. Shannon is a senior program manager in federal IT, an instructor, and a small business owner, so this show sits on both sides of the desk.

Here is how it works. Every week we take three stories and give each one a Hype Check. There are four possible verdicts. Real. Real, but early. Overhyped. And too soon to tell. When something really is a risk, we will say so. Then we will talk about what to do.

Story one. On October twenty second, economists at Goldman Sachs published a new estimate, reported by Fortune. It says up to 9 percent of workers could be displaced during the shift to AI. That is roughly one worker in eleven. Slower hiring could weigh on employment by another 3 percent. Goldman economist Joseph Briggs said on a podcast in June that this share would be about 15 million workers over a decade.

That number will travel fast, so let's run the Hype Check. The verdict is too soon to tell. This is a forecast from a model, not a count of jobs already lost. It assumes AI eventually lifts productivity by 15 percent. It also counts people who lose a job, even if they find a new one quickly. Briggs himself says unemployment would rise by less than one point a year. He also says 9 percent could turn out to be low. So the honest answer is that nobody knows yet.

Here is what we can measure today. The outplacement firm Challenger, Gray and Christmas tracks the reasons employers give for job cuts. Through September, employers cited AI for more than 120 thousand announced cuts. That is about 21 percent of all cuts, and it is the top reason this year. But in September alone, AI was cited for just under 4 thousand cuts, about 9 percent of the month's total. Keep in mind that these are the reasons employers give, which is not proof of the cause.

So why not relax? Goldman's view is that automation job losses tend to stay limited when the economy is healthy, then speed up in a slowdown. Quiet months may not tell us much. And a second Goldman note, reported by Fortune on October thirtieth, says workers who lost AI exposed jobs took about a month longer to find new work than other displaced workers. They were also about 10 percentage points more likely to still be unemployed a year later. Goldman says it is still too early to judge some of these effects, and it had to estimate which losses were tied to AI. But this part deserves your attention.

What does it mean for you? If you are a worker, do not plan your life around one in eleven. Plan around a slower job search. The calm months are the time to update your resume and learn how AI touches your work. If you run a business, this forecast is not a reason to cut. Goldman says only 23 percent of American business locations have built AI into their regular work. Most of the productivity is still a promise, so hold off on any staffing decision that assumes it has already arrived.

Story two. You may have seen the word jobpocalypse attached to new college graduates. On October seventh, the UCLA Anderson Review reported on research by Robert Fairlie and Jane Wu. They looked at unemployment for graduates aged 22 to 25, across five summers. This past summer it averaged 7.3 percent. That was higher than three of the four earlier summers and lower than the 7.8 percent of 2024, but the researchers say it was not significantly higher than past summers. Graduates in more AI exposed occupations did look slightly worse off, but the gap was small and could be chance.

The verdict is overhyped. The headline is bigger than the facts. But hold on, because that is not the whole story. Stanford researchers looked at payroll data through June of this year. Workers aged 22 to 25 in the most AI exposed jobs are employed at levels 19 percent below where they would be if they had kept pace with less exposed peers. Experienced workers show no comparable gap. And the gap has been growing. Stanford calls this an early signal rather than proof that AI is the cause, and it says it does not see widespread job loss across the economy. The pattern weakens when researchers control for education, and some of the trends began before generative AI.

Those two findings can both be true. One asks whether graduates overall are out of work. The other asks how many young people are getting hired into the most exposed jobs. The researchers say the gap comes mainly from less hiring, not more firing. So the door is not slammed shut. In some jobs, it has narrowed.

If you are early in your career, aim for work where AI helps a person do the job, not work where it does the whole task. Stanford found employment is flat or rising in jobs where AI supports people, especially for experienced workers. Be ready to show how you check AI's work. If you lead a team, here is a question worth asking. If we stop hiring beginners, who becomes our experienced people in five years?

Story three. On October twenty ninth, the Associated Press reported on what some people now call the AI boomerang. That is when a company cuts workers because of AI, then brings people back. The AP names two examples. The payments company Klarna started recruiting human customer service reps again after its chief executive told Bloomberg it had gone too far in relying on a chatbot. And Ford said in June that it hired or promoted 350 veteran engineers after AI quality control tools did not perform as expected. Ford also credits AI, together with those engineers, for its quality gains.

The verdict is real, but early. The examples are real. But the AP says it is unclear how common this is, and companies do not always announce it. Meanwhile, the AP notes that Microsoft, Barclays, and Kroger are among the companies that have recently announced AI related cuts. There is a forecast too. The research firm Gartner predicts half of the companies that cut customer service staff because of AI will rehire for similar work by 2027. That is a prediction, so hold it loosely.

If a former employer calls you, you have more to discuss than the start date. Experts in the AP story say returning workers are asking about pay, training, and how AI could affect staffing later. One management professor suggests asking for a one year renewable contract, with a buyout if the job ends early. Whether you can get that depends on your situation. And you do not owe anyone a yes.

If you run a business, the lesson is about the work nobody wrote down. A Forrester analyst told the AP that AI often misses the unwritten work people do, like reading the room and building trust. Before you cut a role, map what that person actually does. And decide who is responsible when the tool gets something wrong.

Our Word of the Week is displacement. Displacement means losing a job because the work changed or went away, whether or not you find a new one soon. It is different from unemployment. If a bank teller's branch closes and they start at a credit union a month later, they were displaced, but only out of work for a short time. So next time you hear a displacement number, ask how many of those people landed somewhere.

Now for our Job Reality Check. Each time, we take one job and look at it honestly. Today it is customer service.

Here is the hard part first. The Bureau of Labor Statistics counts about 2.7 million customer service representatives, with median pay of 44,770 dollars. It projects those jobs will shrink about 5 percent over the ten years to 2035, as self service tools and automation take on simple requests. And in August, Goldman Sachs reported that American call center employment was 39 percent below its historical trend. That gap may not all be AI, but this field is feeling real pressure.

Now the other side. The same government projection expects about 289 thousand openings a year, all from replacing people who retire or move to other work. Those jobs are not vanishing overnight. And a Gartner survey from October of last year found only 20 percent of customer service leaders had reduced staff because of AI, while more than half kept staffing steady as they handled more customers. That survey is a year old, and a newer one shows more movement. In April, Gartner reported that 31 percent of service and support leaders had made or were planning AI driven frontline layoffs through the first quarter of next year. Gartner also found that only about one quarter of the customer service AI uses it evaluated paid off. Another quarter lost money, and 42 percent had unclear value.

So what is moving? The simple, repeatable requests. Password resets. Where is my order. Basic return questions. Those are going to bots and self service, and our read is that they are unlikely to come back.

What stays with people? The customer who is upset. The billing dispute that crosses three departments. The problem that does not fit the script. The judgment call about when to bend a rule.

One thing to watch. Simple tickets have long been where new reps learn the job. If those disappear, the first rung of this ladder could get harder to reach, which fits what we heard in story two.

So how do you become the person running the new workflow? A Forrester analyst told the trade outlet CX Dive that frontline reps may become what he called bot unblockers, people who oversee groups of AI agents. Start now. Notice where your company's bot fails. Keep a list of those failures and how you fixed them. Volunteer to improve the help articles. The person who can explain why the bot got it wrong becomes hard to replace.

That brings us to Skill Up. This week's skill is writing the manual for your own job. It connects to story three and to today's look at customer service. The unwritten work is what companies keep missing, and it is the work that protects you. You are going to use an AI assistant to get it on paper, and you can do it in 15 minutes with any free assistant.

First, set a timer and list everything you did last week, especially the things nobody assigned. Leave out customer names, company data, and anything confidential. Second, paste your list in and ask the assistant to sort it into three groups. Tasks a tool could do. Tasks a tool could help with. And tasks that need a person. Third, read the sorting critically, because the assistant is guessing. Move anything it got wrong, and write one sentence on why. That sentence is the valuable part. Fourth, pick one task and have the assistant draft simple steps for it. Check every step yourself.

Now you have something most people do not have, a clear picture of where you add value. In an interview, you can say this: I documented my team's unwritten work with help from an AI assistant, checked it by hand, and used it to show which tasks should stay with people.

One last thing about responsible use. Check what the AI gives you, keep confidential information out of it, and follow your employer's rules on AI tools.

That is it for episode one. You do not have to follow everything, and you do not have to be afraid of every number. Ask whether it is a fact or a forecast, and ask what you can do this week. This has been AI Fear Check. Take care of yourself, and we will see you next week.