Why do leaders still get communicating about AI so wrong?

Exactly one year ago, I wrote in this column about Duolingo. Its CEO, Luis von Ahn, had taken to LinkedIn to address employee concerns after announcing an “AI-first” strategy, including plans to gradually stop using contractors for work AI could do.

We’re here again. This time, it’s Standard Chartered CEO Bill Winter. When addressing investors about the bank’s AI-driven restructuring, he described some roles as “lower-value human capital”. He later apologised (twice) on LinkedIn, acknowledging his “choice of words” had caused upset and reiterating the organisation’s commitment to supporting colleagues through change.

Having spent part of my career in investment banks, I’ve always struggled with the idea that “back office” roles are somehow lower value. Banks simply don’t function without them. They manage all the operational and admin functions that keep the bank running smoothly. I can only imagine how many hardworking colleagues must have felt hearing those remarks.

Why do some senior leaders still struggle to talk about AI in a way that balances strategy, humanity and trust?

Partly, it’s a language problem.

Employees pay close attention to how leaders talk about people, especially in moments of uncertainty. Terms that might feel routine in financial or strategic discussions can sound deeply dehumanising when they reach a wider audience. When leaders talk externally about efficiency, cost reduction or workforce change before they’ve built an internal narrative about people, skills and opportunity, it contributes to a climate of fear. And once employees feel that the “real story” is being told externally to investors it makes it much harder for your internal messages to be credible.

The recently released IC Index reinforces just how much work needs to be done:

Employees are not clear on how AI will be used in their organisations and how they are expected to use it in their job.

• Only 36% say leaders have clearly explained how AI will be used
• Fewer than one in three understand what AI means for their own role

Employees are realistic. They understand that roles will evolve and some tasks will disappear. But what they are looking for is evidence that leadership has thought about the human implications This is where internal communications has a critical role. In helping leaders communicate in a way that is both commercially credible and human.

In practice, that means moving from broad statements about AI to a clear, consistent narrative that people can understand and engage with.

Here are five practical ways to do that:

1. Anchor your narrative in why AI matters to your organisation. What problems are you trying to solve? What outcomes are you aiming for? If you don’t lead with purpose, people will assume the story is about cost before anything else.
2. Make it explicit how people feature in your plans. What skills will be needed? Where are the opportunities? What support will be available?
3. Separate what’s decided, what’s in testing, and what’s still evolving. Avoid overpromising and be clear about where things are still uncertain.
4. Translate AI into day-to-day impact. Focus on the practical: what might change, what might stay the same, and what they should start doing differently.
5. Equip leaders to communicate with humanity by explaining things simply, acknowledge uncertainty, and connect decisions to people.

The reality is that most organisations don’t have all the answers on AI yet, and they may not for some time The organisations that will navigate this well are the ones that communicate early, honestly and with people at the centre of the story. Because in the absence of a clear narrative, employees will create their own. And once that happens, rebuilding trust is a much harder job