Key Takeaways
- Silence from leadership breeds anxiety. Employees fill the void, and not with optimism
- Behavioral data tells you who will thrive in a restructured role, before you find out the hard way
- Your employer brand is what employees say about you when you’re not in the room
- HR needs a seat at the table before the restructuring, not just to clean up after it
Meta’s CTO Andrew Bosworth got on an internal call recently and said morale is probably near the worst it’s ever been in his 20 years at the company. He compared it to Cambridge Analytica. His exact words: “The vibes are off.”
What’s striking isn’t just that he said it. It’s that he also admitted leadership “did an atrocious job explaining the vision.” That’s a rare moment of executive candor, and it points directly at the real problem.
This isn’t primarily a layoff story. It’s a change management story. And HR teams across every industry should be paying attention, because the same dynamic is coming for them.
What actually happened at Meta
Between early 2025 and May 2026, Meta ran multiple rounds of cuts, around 8,000 roles gone in May alone, about 10 percent of the workforce, while simultaneously building out an aggressive AI-first strategy.
The flashpoint was a unit called Applied AI, stood up in March and immediately stocked with 6,500 engineers and product managers who had little say in the move. Engineers who were previously building products used by billions of people found themselves writing coding puzzles and data labeling tasks to train AI models. The work felt menial. The career path was unclear. One person described it to Wired as “literally the gulag.” Another called it “soul-crushing.”
Meanwhile, Meta was installing software on employee laptops that logged keystrokes and screenshots to train AI, with no opt-out. More than 1,000 employees signed a petition against it. And median total comp dropped from $417K to $388K, right after one of the company’s most profitable quarters ever.
The anger makes sense. And it’s worth being precise about why.
The real driver isn’t just job cuts
Layoffs damage morale. That’s not surprising. What’s more interesting here is that the damage is specifically concentrated in the Applied AI group, people who still have jobs, who are well-paid by any standard, but who feel trapped and disengaged.
That tells you something important: this isn’t primarily about job security. It’s about meaning and autonomy.
Research consistently shows that when people lose ownership of meaningful work, when the thing they were hired to do gets replaced by tasks that don’t connect to their skills or motivations, engagement breaks down fast. That’s not a hypothesis. It’s foundational to how behavioral science understands what drives people at work. At PI, we’d describe it in terms of a person’s drives: their need for autonomy, for ownership, for doing work that uses the capabilities they’ve built. Strip those away without explanation, and you don’t just get frustration. You get the kind of deep disengagement that perks and all-hands meetings can’t touch.
Applied AI, as Meta has structured it, is directive, top-down, and tied to a broad organizational mandate. That works really well for some people. It actively repels others. And when you move 6,500 people into that environment without any behavioral consideration, without understanding who will thrive there and who will shut down, you get exactly what happened: petitions, Blind threads, and a CTO on a public call comparing the mood to a corporate scandal.
The problem isn’t the AI pivot. The problem is treating a behaviorally diverse workforce as if they’re all wired the same way.
What this means for AI rollouts
Here’s the thing about AI-driven restructuring: it’s going to keep happening. The companies that handle it badly will leave a trail of disengaged employees and damaged employer brands. The ones that handle it well will do one thing differently, they’ll treat the human side of the transition as seriously as the technical side.
That means a few concrete things.
First, people need to understand why the work matters. Not in a vague “this is the future” way, but specifically: how does writing this evaluation task connect to the product we’re building, and what does that mean for my career? The research on AI adoption is clear that employees don’t resist change because they’re inflexible. They resist when the vision is unclear and the personal implications are left to their imagination.
Second, not everyone experiences this kind of transition the same way. Some people adapt quickly when given clear direction. Others need to understand the bigger picture before they can commit. Some are energized by new problems; others feel destabilized when their area of expertise suddenly shifts under them. A blanket mandate, “you’re in Applied AI now”, doesn’t account for any of that. Understanding how your people are wired tells you who needs what during a hard transition, and that’s where behavioral data becomes genuinely useful, not as a hiring filter, but as a management tool in exactly these moments.
Third, trust is fragile and slow to rebuild. The keylogger story matters not because of what it does, but because of what it signals: we don’t fully trust you, and you don’t have a say in how we observe your work. That kind of signal compounds everything else. Once employees feel surveilled rather than supported, engagement data becomes noise, people perform compliance, not commitment.
Where HR comes in
HR’s job in an AI transformation isn’t to absorb the fallout after decisions are made. It’s to be in the room when the decisions are being made.
That’s a harder sell than it sounds. When a company is pivoting to AI at Meta’s speed, the pressure to move fast is real. The analysis is financial and strategic. The human cost gets managed after the fact, usually with a town hall and a new perk.
But the Meta situation shows exactly what that approach costs. The employer brand damage from a CTO publicly comparing morale to a corporate scandal will outlast the restructuring. The engineers who leave will talk. The ones who stay but disengage will be harder to re-motivate than they were to hire.
HR teams who want to avoid this have to push for behavioral intelligence to be part of restructuring decisions upfront, not just which roles to cut or consolidate, but what the people in those roles need to stay engaged, and what the transition plan looks like at the individual level. That’s a fundamentally different ask than “run a change management communication plan.” It means treating your people as individuals with different behavioral profiles, not as a workforce to be administered through policy.
Meta’s situation isn’t unique. It’s a preview. The AI restructuring wave is moving through every industry right now, and most companies are running the same playbook: move fast, communicate after, manage the fallout. Some of them will get lucky. Most will end up with their own version of “the vibes are off.”
Meta’s initial response to the crisis, for what it’s worth, was better snacks and a manager headcount cap. That’s not nothing. But if the article you just read made any sense, you already know it’s not enough.
The antidote isn’t more communication. It’s understanding your people well enough to know what the transition means for each of them before you ask them to make it.
