Meta CEO Mark Zuckerberg informed his workers that the company had made mistakes as it transformed its workforce around AI, according to reports from Reuters.

This is no minor acknowledgment, as Meta had invested heavily in shifting the company towards becoming AI-based. But clearly this process has not gone quite as smoothly as the Silicon Valley company would want.

Meta’s AI push is getting uncomfortable

Zuckerberg reportedly told staff that Meta has “made mistakes” during its AI transformation and will likely make more because the changes are complex. He also said he wants to provide more stability and does not expect more company-wide layoffs this year.

That matters because Meta isn’t just adding a few AI tools. It’s trying to rebuild how the company works.

According to Reuters, Meta has invested heavily in AI, cut around 10% of its global workforce, and reassigned about 7,000 employees into AI-related roles.

That kind of shift changes people’s jobs overnight.

For employees, the problem isn’t only whether AI will replace work. It’s whether companies can explain what the new work actually is.

What went wrong inside Meta?

The pressure appears to centre on Meta’s Applied AI Engineering unit, where workers reportedly raised concerns about broad management spans and unclear roles. Reuters says Meta now plans to ease back on those demands and invest more in offsite events, corporate gatherings, and a July hackathon.

Here’s the simpler version:

Issue	What it means
Reassigned workers	Staff moved into AI roles they may not have chosen
Management pressure	Managers handled wider teams and more complexity
Morale problems	Employees questioned the purpose of some new work
Stability promise	Zuckerberg says no further company-wide layoffs are expected this year

WIRED also reported frustration inside Meta’s new AI unit, with some employees describing the work as low-value and demoralising. The report said internal tension grew after layoffs and forced moves into AI-related teams.

That’s the human side of the AI boom.

Meta is still spending big on AI

The awkward part is that Meta can’t simply slow down.

Reuters reported that Meta raised its annual capital expenditure forecast to between $125 billion and $145 billion, showing how serious the company remains about AI infrastructure.

That money goes into data centres, chips, servers, and the systems needed to train and run large AI models. It’s expensive before it becomes profitable.

And Meta isn’t alone. Big Tech companies are racing to build AI infrastructure because they don’t want to fall behind OpenAI, Google, Microsoft, Amazon, or Anthropic.

But there’s a catch.

The AI race needs money, talent, and speed. Workers need clarity, trust, and time. Those two realities often clash.

Why South African companies should care

This story may feel far away from Johannesburg, Cape Town, or Durban. It isn’t.

South African companies are also adding AI to customer service, marketing, banking, HR, and internal operations. We’re already seeing AI tools shape how businesses talk to customers on platforms like WhatsApp. For more context, read our breakdown of Meta’s WhatsApp AI agents and what they could mean for South African businesses.

The lesson from Meta is simple: AI transformation cannot only be a tech strategy.

It’s also a people strategy.

If a company rolls out AI without explaining how jobs will change, workers will fill the silence with fear. If it moves people into AI roles without training, it creates confusion. If it tracks work behaviour to train AI systems, it also creates privacy and trust questions.

That matters in South Africa because workplace data is not just “company data.” Under POPIA, businesses must think carefully about how they collect, process, and protect personal information.

So when global tech giants experiment with AI-driven work models, local companies should watch closely. Not because they must copy them, but because they can learn from the mess.

The bigger question for Meta

Zuckerberg’s admission doesn’t mean Meta is backing away from AI.

It means the company knows its internal shift has created friction. The next test is whether Meta can turn that admission into better execution, not just better messaging.

AI may help companies move faster. But if workers feel drafted into unclear roles, watched by new systems, or pushed aside for automation, the transformation becomes harder to sell.

Meta wants to build the future of AI.

But first, it has to convince its own people that they still have a meaningful place in it.

FAQs

What did Zuckerberg say about Meta’s AI shift?

Zuckerberg reportedly told employees that Meta has made mistakes during its AI workforce transformation. He said the company will probably make more mistakes because the shift is complex and fast-moving. The message shows Meta knows its AI push has created real tension inside the company.

Is Meta planning more layoffs?

According to Reuters, Zuckerberg said he does not expect more company-wide layoffs this year. That gives employees some short-term reassurance after Meta’s latest restructuring. Still, teams may continue to change as the company keeps reorganising around AI.

Why does this matter in South Africa?

South African companies are also bringing AI into everyday work, from customer service to marketing and banking. Meta’s situation shows what can go wrong when AI changes jobs faster than workers can adapt. For local businesses, the lesson is clear: AI adoption needs training, transparency, and strong privacy safeguards.