Are Some Companies Citing AI to Justify Layoffs Before They Can Prove AI Actually Replaced the Work?

Three major tech companies cut workers and point to AI as the reason within the same three-week window. Cloudflare frames its 1,100-person reduction around the agentic AI era. Amazon and Meta make similar moves with similar language. The corporate narrative shifts from "we over-hired" to "AI is replacing the work," but the evidence that AI actually replaces those specific roles remains thin. For most organizations, the layoff announcement runs well ahead of any measurable productivity proof.

Joseph Mugare, Data Analyst at Colgate-Palmolive, has spent over nine years across data science, full-stack development, and cloud infrastructure roles spanning the United States and Kenya. His work covers data analysis, reporting automation, and market intelligence for one of the world's largest consumer goods companies. He holds AWS certifications and has completed advanced coursework in Python, SQL, and data analytics through Stanford, Harvard, and MIT.

"At the moment, we are still early. There's no decision that can clearly say AI has replaced these jobs. We're still learning," says Mugare.

The Trainers Lose Their Jobs First

Mugare notes a pattern in the recent layoffs that makes the narrative harder to accept at face value. Many of the workers cut are the same people who built and trained the AI systems being credited with replacing their roles. "The people who lost their jobs were the trainers, the ones who trained the AI. They trained something that was going to take their jobs away."

The companies frame the cuts around output efficiency, arguing that AI produces faster and more consistent work than the employees it replaces. But Mugare questions whether those claims hold up to scrutiny when the industry is still only three years into widespread generative AI adoption.

He sees the dynamic as a broader structural issue rather than a technology-driven inevitability. When asked whether workforce reductions reflect genuine AI capability or something else, his answer is direct. "I can say it's something to do with capitalism. Different companies are looking at efficiency and output, and AI gave them the narrative."

Upskilling, Not Replacement

From his own work as a data analyst, Mugare describes how AI is already changing daily workflows without eliminating the role itself. AI-enabled Excel can generate reports, automate documentation, and accelerate analysis.

"In the near future, you just need to learn a prompt, and you can generate a report." But rather than viewing that as a threat, Mugare frames the response as a skills question. "It's not just about layoffs. It's about people asking, what else can I learn so that AI cannot replace me? What else am I adding?"

He argues that organizations still trying to draw a clean one-to-one line between AI adoption and headcount reduction are getting ahead of themselves. The tools are changing the work, but the boundary between what AI can fully execute and what still requires human judgment is far from settled.

"AI is still getting better. So the boundary between human work and machine work is not clear yet." The risk he sees is losing what he calls the human touch: the ability to think generatively, bring diverse ideas, and make judgment calls that a prompt cannot replicate.

Emerging Markets Tell a Different Story

Mugare, who has worked across both the US and Kenyan markets, points out that the AI employment picture looks very different in emerging economies. In Africa, AI is actively creating jobs in data annotation, model training, and application development.

"Right now there are people building different applications at a much faster rate, maybe for agriculture. It's creating employment for young people. People can create apps in a matter of minutes, sell them, and make a living."

But he acknowledges the fragility of that employment cycle. Data annotation and training roles are inherently temporary. Once the models are trained, the demand for those workers shrinks. The same AI capabilities creating jobs today can erode them later.

The broader question, Mugare argues, is whether companies can honestly claim AI has replaced specific work or whether the layoff framing is simply more convenient than admitting the gains are still unproven. "We are around three years since we started having ChatGPT and the rest. I think it's still early. It's still early to know."