Everywhere we turn, we hear about the transformative potential of generative AI. A flood of toolkits and opinion pieces promises faster work, fewer dull tasks and more time for high-value thinking. Social feeds are saturated with AI hacks for boosting productivity and automating anything that feels inefficient.

Let me be clear: GenAI really can help us work faster and scale more easily. But when we focus solely on what AI can do, we risk losing sight of what we may be giving up. Leaders must start asking a deeper question: Where does real value come from in our work? If we don't, we risk destroying more value than we create.

Where value lies

Writing this article involved several false starts, abandoned outlines and a “scraps” document much longer than the final version. Out of curiosity, I asked a GenAI tool to draft it from a few prompts. In 10 seconds, I had something that was in fact very good. The speed and efficiency were impressive. But what did I lose in skipping the messy process?

While there is no denying the productivity, innovation and creativity benefits of Gen AI, they aren’t the only kind of value we get from our work. Here are five dimensions of value leaders often overlook when they rush to hand over tasks to GenAI.

1. Knowledge and learning

Many work tasks teach us something indirectly. Struggling to remember a word in a second language helps cement it in memory. Trying multiple solutions to a tough problem often leads to accidental insights (as was the case with penicillin, Coca-Cola and smoke detectors). Synthesising information improves our understanding.

Letting GenAI translate, solve or summarise may be faster and sometimes more accurate. But when we skip the process, we lose the learning. As the Confucian philosopher Xun Kuang reportedly said, “Tell me and I forget. Teach me and I remember. Involve me and I learn.”

2. Skill building

We get better by doing. Writing and revising sharpens our editing. Debugging code deepens our programming skills. Grappling with writer’s block builds resilience. Relying on GenAI to clean up our drafts or propose options gives quick results, but over time it erodes the very skills we need to keep developing.

3. Connection and collaboration

Traditionally, complex work has been done in teams. GenAI now gives individuals the computing power to do much more alone. That means fewer shared moments of problem-solving, and fewer opportunities to build trust, understanding and camaraderie.

These interpersonal dynamics aren’t just nice-to-haves. They’re central to organisational health. Without them, performance suffers, and so does well-being. Research shows we’re already seeing signs of growing workplace isolation.

4. Active engagement

When we’re fully mentally present in our work, we do better and feel better. GenAI can diminish that presence. For example, when it summarises a report or meeting transcript for us, we may no longer process the material ourselves.

Summarising is not just about extracting key points. It’s an act of framing and interpreting meaning. That process deepens our understanding and increases our engagement. 

While it’s easy to claim that we’ll add our own input later, ask yourself how often you really do that. We’re all busy, and it’s just too easy to skim, nod and say, “Yeah, I think this is good enough,” without engaging deeply enough to recognise what’s missing.

5. Voice and identity

Every leader has a unique style, reflected in how they write, speak and solve problems. These traits shape their professional identity. But when AI writes your memos or builds your presentations, your personal tone and voice risk getting flattened.

Large language models are built to converge, producing broadly acceptable answers. So when two leaders ask the same model the same question, they get similar responses. Over time, that can dilute not only individual identity but organisational distinctiveness.

Short- and long-term effects

These costs are not always immediate or obvious. Comparing the productivity gains from GenAI to lost learning or reduced trust is like comparing apples to motor oil. The dimensions differ, and so do the effects.

In work I’ve done with Harvard Business School professor Amy Edmondson on the employee value proposition, we highlight two axes for leaders to consider: individual vs. collective benefit, and short- vs. long-term impact. A GenAI application might increase team output now but hinder individual development later. Or it might save time today but reduce engagement over time.

Humans aren’t naturally good at long-term thinking. But with GenAI, we need to stretch our time horizon and look beyond immediate productivity.

How to run an AI value audit

So how should leaders assess where GenAI fits? Like any managerial decision, it requires a clear-eyed review of costs and benefits. I suggest conducting an AI value audit. Here’s how.

Step 1: Identify value types. What value does a task generate? Think of speed, accuracy, insight, learning, connection or voice. Not every task delivers every type of value, but leaders must think beyond output alone.

Step 2: Prioritise and balance. Consider trade-offs. AI-powered meeting notes may boost efficiency but lower engagement. If that’s a concern, you may decide to automate only low-stakes conversations or alternate between AI and live note-taking. If GenAI replaces a brainstorming session, create other chances for team bonding.

Step 3: Revisit your assumptions. Treat AI decisions like perishable goods. Review them regularly, don’t discard without reason, but don’t assume they’re still fresh. Good scientists form hypotheses, test them, gather feedback and adjust accordingly. You should do the same.

It’s not just up to leaders

One final note: These decisions aren’t solely in leadership’s hands. GenAI is accessible, cheap and everywhere. Employees are already making choices about when and how to use it.

So look at your metrics. If your KPIs measure only speed and output, people will naturally lean on AI. That might be fine, but only if that’s truly all that matters. If it isn’t, rethink what you reward.

And most importantly, talk to your people. Help them understand the kinds of value GenAI enables – and the kinds it can quietly undermine. That’s the only way to ensure they’re making choices that serve both their growth and your organisation’s future.
 

This is an adaptation from an article originally published in the Harvard Business Review.