There is a word that almost never shows up in an efficiency deck: trust. And it is the exact word that explains why Meta, after years as the most imitated management model in Silicon Valley, just admitted its workplace morale is the worst it has ever been.

Here is the thesis up front, because it is the part worth keeping: efficiency that skips trust is not free. It feels free for a quarter or two while the numbers climb. Then the bill arrives, and you pay it with the same people who produced those numbers. What happened to Meta is not a culture problem unique to Menlo Park. It is what happens any time a transformation, especially an AI one, gets pushed top-down with the people left outside the room.

What Meta just admitted

Business Insider titled it with surgical precision: “Meta culpa”. Aki Ito’s piece reports something that would have been unthinkable a year ago: the company’s own leaders, out loud, acknowledging a collapse in morale and trust.

The clearest was CTO Andrew Bosworth. In an internal meeting on June 2 he said morale is “probably one of the worst it’s ever been.” Then he reached for a comparison: Cambridge Analytica, the 2016 scandal that nearly broke the company. When the second-in-command picks that analogy, facing inward, it says more than any survey.

He was not alone. CPO Chris Cox spoke of the “insanity of this company” and a “difficult” and “brutal” environment. Zuckerberg himself conceded, “we’ve made mistakes.” A Harvard Business School professor summed up the damage in the same coverage: Meta has “almost systematically destroyed trust” and is now trying to dig itself out of the hole it dug.

The backdrop is the May numbers. Meta cut around 8,000 people, roughly 10% of staff, and reassigned another 7,000 to train AI models. As Futurism documented, the mix of mass layoffs and a “the agent does the work, you supervise it” message did not produce the sharp, lean org the playbook promised. It produced burned-out people.

Why efficiency without trust gets charged back

It helps to separate two things that get blurred constantly. Cutting headcount is not, by itself, what breaks a company. Restructurings exist and are sometimes necessary. What breaks a company is treating efficiency as free, when it actually runs on an invisible asset called trust.

Trust is what makes someone load the full context of a problem without being asked, flag an error before it detonates, stay an extra hour because the result matters to them. None of that shows up on a dashboard. All of it is what produces real efficiency, the kind you do not have to police.

When you drain it, the person stays in their chair but stops doing those things. They do the minimum that can be verified. The trap is that no productivity metric warns you in time, because trust is not measured, it is spent quietly. By the time it shows up in turnover or a project that collapses, you already have months of damage on the books.

Meta is the extreme case because it had record margins to sustain the pressure longer. A mid-sized company does not. Copy the ruthless-efficiency script without Meta’s balance sheet, and the bill arrives sooner and hurts more.

I wrote about the mechanism in April. Now the bill is here.

In April I covered the other side of this same story, when Meta started recording employees to train their replacements. Back then the subject was the mechanics: the tool capturing every keystroke, the clause set to creep into SaaS contracts. The warning pointed outward, at what you might sign without noticing.

This is the inward half. The same decision that looked like a cost play in April produced, two months later, the worst morale in the company’s history. These are not two separate news items. They are cause and effect, six weeks apart, between “your work starts being recorded” and “your team stopped trusting you.”

It ties to something I keep arguing about how expert certainty blocks AI adoption: bad AI decisions almost never come from the technology. They come from imposing a belief, here “efficiency always pays,” without testing it against the real data of how your people actually work and react.

What we do differently at IQ Source

Meta’s lesson is not “don’t adopt AI.” It is don’t adopt it as a cut dressed up as strategy. The difference between a transformation that adds and one that destroys trust is not how much AI you deploy. It is how you deploy it and who you bring into the change.

That is why AI Maestro does not start by installing an agent or announcing who it will replace. It starts with two months of discovery, where we map how your processes actually work and where AI adds real value. That map does not come from the opinion of whoever outranks the room. It comes from process data, which is why the people who do the work help build it instead of finding out after the decision is made.

That difference is what separates an adoption a team pushes from one a team endures. When AI comes in to take the routine half off people and leave them the judgment half, the team defends it. When it comes in as a threat, the team protects itself, and that is where the collapse Meta just admitted begins.

Tomorrow I publish the exact opposite of this story: IKEA, facing the same technology, chose to reskill 8,500 people instead of firing them. Same AI, opposite result. The variable was never the model. It was leadership.

If you are about to bring AI into your company and your plan starts with how many people you will cut, you do not have an AI strategy. You have Meta’s playbook. And you just watched how it ends.

Adopt AI without burning your team’s trust