The number traveling around social media - 205,000 US jobs lost to AI - doesn't hold up. The verified count from Challenger, Gray & Christmas tells a different, more useful story: US employers cited AI in 112,713 announced job cuts through July 2026, more than double the 54,836 AI-cited cuts tracked in all of 2025.
AI led all stated reasons for job cuts in July for the fifth month in a row, with 10,970 cuts that month - 33% of all announced cuts. But here's the part that gets lost: across the first seven months of 2026, US employers announced 477,033 cuts overall, down 41% from the same period in 2025. The labor market isn't collapsing in one direction. AI is becoming a more common explanation inside a smaller layoff total.
That distinction matters for your job. When companies name AI, investors hear discipline and margin expansion. Workers hear something blunter: the task you were paid to do can now be done with fewer people around it. Both readings can be true at once.
Where the cuts are landing
Challenger's July report put technology at the center, with 149,023 announced layoffs through July, up 67% from the same point in 2025. The firm also made a careful distinction that gets lost in louder summaries: some cuts are directly attributed to AI, while others sit in a separate technology-update bucket when AI is only implied. That difference decides whether a layoff is counted as automation or ordinary restructuring with better branding.
Salesforce is the cleanest example. CEO Marc Benioff said on The Logan Bartlett Show that Salesforce reduced support headcount from about 9,000 to about 5,000 as AI agents took over more service work. The San Francisco Chronicle reported that Benioff said AI now handles about half of Salesforce customer conversations.
Amazon is messier. The New York Times reported that Amazon cut 16,000 corporate employees in January 2026 after cutting 14,000 in October 2025, while pouring money into AI data centers. Andy Jassy had told employees in June 2025 that AI would mean fewer corporate workers over time, but the Times also reported that he later framed the layoffs less around AI and more around reducing bureaucracy. Don't flatten that into a clean AI-replacement story. Amazon didn't.
The bottom rung is disappearing
Duolingo gives you a clearer view of the work most exposed. The Washington Post reported in January 2024 that Duolingo cut about 10% of its contractors at the end of 2023, including people who wrote lessons and translation variants, while using more AI tools to create language content. Duolingo's own strategy materials say the company published 20,500 skills across its language courses in Q1 2026, up from 7,100 per quarter in 2025 and 1,800 per quarter in 2024. More course output, fewer old content-production paths.
If you're early in your career, that's the part to watch. Customer support, content work, translation variants, data operations, basic code maintenance - those were the jobs where people learned the machinery of a company. Take them away too quickly and you don't just cut cost. You cut the training ground.
Klarna is the cautionary tale. Bloomberg reported in May 2025 that CEO Sebastian Siemiatkowski said the Swedish fintech's cost-focused customer-service push had gone too far and produced lower quality, and that Klarna was testing a new group of remote human support workers so customers could still reach a person. This came after Klarna had said its AI assistant was doing work equal to hundreds of agents. AI can cut handle times and payroll, then still leave the company paying for the human repair work later.
What this means for people in customer support
The buying signal for AI tools is real. Companies are willing to put AI into official layoff language now, and the tools are getting better at ticket triage, draft replies, and routine resolution. But the Klarna example shows the ceiling on pure automation. Support work has edge cases that AI still can't handle - the angry customer, the confusing billing problem, the product issue that doesn't match any training data.
Your best move is to learn how these systems actually work rather than wait to be managed by one. Understanding what AI can and can't do in a support workflow - and being able to operate the tools yourself - is the difference between being replaced by the technology and being the person who makes it work. That's the skill set that survives the current round of cuts, and it's also the path into the roles that companies still need: supervising the AI, handling the escalations, and fixing the quality problems that Klarna ran into.
Your membership also unlocks:
