One of the more misleading assumptions about AI is that it will simplify work in a clean, predictable way.

Sometimes it does. But often it adds a second layer of complexity that leaders did not anticipate.

The January 22, 2025 Harvard Business Review article warns that AI is not merely changing productivity. It is also
changing talent risk, employee expectations, and the hidden pressures inside organizations.

In other words, AI does not replace the challenge of leading people. It often intensifies it. 

That matters because emerging talent risks are not always obvious at first. A person may appear more productive with AI support while still struggling with judgment, interpersonal maturity, ownership, or adaptability. Negative behavior patterns can also become harder to spot. Friction, avoidance, dependency, quiet resentment, and passive resistance do not disappear when new tools arrive.

In some environments, they actually increase as employees feel disrupted, threatened, or forced into workflows they did not help shape.

That is part of what makes AI adoption more than a technology issue. It is a leadership issue.

The costs are not only financial.

Yes, AI often requires real out-of-pocket investment. But the indirect costs may be even greater: employee friction, workflow redesign, training burdens, trust gaps, and barriers to adoption. Gartner reported that in a July 2024 survey of more than 3,500 employees, one in five said technology adopted by their organization in the previous two years had made their job harder rather than easier or having no impact.

That is a sobering number for any leader assuming that new tools automatically create progress. 

This is why the best organizations will not take a technology-first approach. They will take a human-first approach. Gartner found that when organizations do this, employees are 1.5 times more likely to be high performers and 2.3 times more likely to be highly engaged.

That finding is especially important because engagement and performance rarely improve through pressure alone. They improve when people believe the system is being built with them in mind, not merely imposed on them for efficiency’s sake.

Leaders should remember: we are still leading people.

Tools matter. Workflows matter. Adoption matters.

But in the end, organizations do not succeed because they bought new technology. They succeed because they learned how to help human beings use it wisely, trust it appropriately, and work together without losing
clarity, dignity, or momentum.

That is the real work now.