Every time a senior leader uses AI for internal communication, we go through the same ritual outrage cycle:

“This is dystopian.”
“This proves leaders don’t care.”
“This is the end of authentic communication.”

It’s neat. It’s satisfying. But it’s also mostly wrong.

Take Mark Zuckerberg using Meta’s own AI to communicate with employees as an example. The internet’s read is predictable: CEO outsources talking to staff, inserts robot, job done. Case closed.

Except… not really.

If you run a company the size of Meta, the idea that every piece of communication should be handcrafted, personally delivered, and perfectly consistent across tens of thousands of employees in multiple time zones is fantasy.

At that scale, AI isn’t inherently a cop-out. It’s infrastructure.

Meta’s stated aim here is that employees might feel more connected to Zuckerberg through these interactions.

Connected to what, exactly? Because this isn’t just AI helping a leader communicate more efficiently; it’s AI attempting to simulate the experience of access to that leader.

So, the question isn’t “AI in leadership communication: good or bad?” It’s much less dramatic, but arguably more useful.

The case for AI — and yes, there is one

Before we get purist about this, it’s worth being honest about where AI genuinely earns its keep.

Consistency, for a start. Anyone who’s watched a “single” leadership message get rewritten six times across regions, functions and translation teams knows how quickly meaning drifts. AI can stabilize that.

Reach, too. Not in the “we sent it to everyone” sense, but in the “the right people actually got something usable” sense. AI can help tailor, translate, summarize, and distribute without requiring an army of humans to manually reshape the same message. ​​

AI is ideally placed to do the heavy lifting — getting the right content to the right person, in the format that suits their context. I cover this in depth in Digital Communications at Work, but the short version is simple: relevance is what cuts through noise and gets read.

And then there’s the signaling effect. A CEO using their own company’s technology is making a statement: we believe this works. That matters, internally and externally.

Used like this, AI isn’t a shortcut, but a delivery mechanism.

But scale cuts both ways.

At Meta’s size, every interaction is effectively broadcast. Which means the cost of getting it wrong — of answering the wrong question or answering it in the wrong order — is multiplied, not reduced.

Which means the real issue isn’t whether AI belongs in leadership communication, it’s where using it breaks things.

What employees are actually looking for

There’s a simple model that sits underneath most internal communication, whether we acknowledge it or not.

We communicate because we want people to know, do, or feel something different.

But here’s the bit that gets missed: employees don’t look at comms channels because they’re hoping to feel something. They do it because they need something.

An answer.
A decision.
Clarity on what happens next.

Gallup’s data is blunt on this: fewer than half of employees (just 46%) say they clearly know what’s expected of them at work.

That’s not a tone problem; it’s a clarity problem.

The job of internal communication is to meet that need so clearly and so directly that the emotional response follows on its own.

This is where a lot of AI-generated leadership communication falls over — and not for the reasons people usually give.

It’s not that it “doesn’t sound human enough.” Employees are perfectly capable of tolerating slightly stilted language. Trust me: they’ve been doing it for years.

The problem is sequence.

In high-stakes moments, AI-generated messages tend to lead with feeling — reassurance, vision, values, “we understand this is difficult” — before they’ve answered the basic, functional question the employee came with.

That’s exactly what this kind of “AI leader” is designed to optimise for — the feeling of connection.

The problem is that connection isn’t something you can lead with. It’s something you earn, after you’ve been useful.

And employees clock that immediately.

They opened the message to understand what’s changing. Or whether leadership has actually thought this through. Instead, they get a paragraph that performs care without first demonstrating usefulness.

At that point, the tone doesn’t matter. The trust signal has already been sent — and it’s the wrong one.

And trust isn’t a given to begin with. Gallup finds that just 19% of employees strongly agree they trust their organisation’s leadership.

In that context, how you communicate, and whether it actually answers people’s questions, matters more.

Trust isn’t built on warmth or polish. It’s built on the sense that leadership is paying attention to the things that matter — and responding to them clearly, directly, and honestly.

When leaders do get this right, the effect is dramatic. Gallup finds that when employees strongly agree their leaders communicate clearly and act decisively, 95% say they trust them.

A more useful way to think about it

Most organizations still split communications into “informational” and “relational,” as if those are different categories.

In practice, that’s not how people experience them.

A better question is: What does the employee need from this and are we starting there?

Take a change announcement. It’s often treated as a “relational” moment — reassurance, tone, leadership visibility.

But what employees actually need is brutally practical:

What is changing?
When?
What does it mean for me?

Get that right, and the emotional layer lands. Skip it, and the emotional layer reads as deflection.

This is where AI becomes either useful or actively unhelpful.

Used well, it helps leaders be clearer, faster, more consistent. It reduces friction in getting the right information to the right people.

Used badly, it generates the language of human presence without the substance underneath it.

And that doesn’t just fail to connect. It sharpens the gap.

Employees don’t just feel uncared for. They feel like the absence of care has been automated.

This isn’t happening in isolation. It’s part of a broader shift towards AI “characters” — synthetic influencers, AI customer service personas, virtual interfaces designed to feel like relationships. All aim to simulate the experience of access to someone, at scale, without the cost of the someone.

Leadership communication is just the latest surface this lands on. And employees, who are already being asked to adapt to AI across their working lives, are not going to miss what’s happening.

Start with the problem, not the tool

The organizations that get this wrong tend to follow a familiar pattern.

They adopt AI because it’s available, there’s pressure to be seen to be using it or someone senior has decided it’s the future.

Only afterwards do they ask what they actually needed it for.

If you’re serious about using AI in leadership communication, the starting point is much less glamorous:

What do your employees actually need from you?

A few uncomfortable diagnostic questions help:

• In your last major change, were people asking for more information or more access to leaders?
• When trust dips, is it because people aren’t hearing from leadership or because what they hear isn’t useful?
• Are you solving a reach problem or a relationship problem?

And what’s happening around that communication?

At Meta, employees are simultaneously being pushed to adopt AI tools, assessed on their AI capability, and — inevitably — wondering what that means for their own roles.

That context matters. Because people don’t read leadership messages in a vacuum. Those are not the same thing. AI helps one far more reliably than the other.

Three questions before you hand it over

Before you let AI anywhere near leadership communication, three checks do most of the heavy lifting:

1. What did employees come here looking for?
   If you can’t answer that concretely, the message isn’t ready — AI or not.
2. Does the communication meet that need first?
   Not eventually. Not after three paragraphs of tone-setting. First.
3. Is real human attention visible somewhere in this?
   Not necessarily in every sentence, but in whether the response takes the situation seriously.

That last one is the line AI can’t cross. And the one that matters most when people are uncertain, frustrated, or waiting to see whether anyone is actually paying attention.

The lesson everyone’s missing

The Zuckerberg example isn’t a warning about AI replacing human leadership.

It’s a reminder to be precise about what you’re delegating.

Leaders who use AI well use it to reduce the overhead — distribution, consistency, accessibility — so they can spend their attention where it actually counts.

That’s not a compromise. That’s what good leadership actually looks like in the age of AI.

Because employees trying to make sense of change aren’t asking whether the message sounds authentic. They’re asking something much simpler: did anyone think about what I actually need here?

Delegate that question to an AI, and you haven’t saved time. You’ve answered it. The answer is no.