Employees are being encouraged to embrace AI, experiment with new tools, and rethink how they work. But there's a question many leaders haven't fully answered.
What does all of this mean for the people doing the work?
A recent Harvard Business Review article by Liz Fosslien, Mollie West Duffy, and Kim Scott explores the mixed feelings employees have about AI. People recognize its potential to make their jobs easier, while also worrying about what happens when the technology can perform more of their responsibilities.
The authors make a compelling case for acknowledging those concerns. We agree, but there's another part of the conversation worth having.
Employees need clear expectations, opportunities to develop their skills, and a better understanding of how AI will affect their roles. Those are decisions leadership has to make.
People have good reasons to feel conflicted
The excitement around AI is understandable. People are using it to work through complex information, research unfamiliar topics, write documents, and handle tasks that used to take hours.
At the same time, they're hearing predictions about jobs being eliminated, teams getting smaller, and companies doing more with fewer people. It's reasonable to have mixed feelings, especially when those predictions concern their own livelihoods.
Anthropic's research offers an interesting perspective. In its study, What 81,000 People Want From AI, researchers gathered interviews from 80,508 people across 159 countries. People described the ways AI was helping them, along with concerns about job security, dependence on technology, and losing skills they had spent years developing. Hope and concern often appeared in the same person.
Source: Anthropic, What 81,000 People Want From AI
Another finding from workplace research deserves attention.
29% of employees surveyed reported behaviors classified as sabotaging their company's AI strategy.
That's from a 2026 survey by Writer and Workplace Intelligence, involving 2,400 executives and employees across the United States, United Kingdom, and Europe. The behaviors varied, from refusing to use mandated tools to using unapproved AI applications. They shouldn't all be treated as equally serious, but the finding points to significant friction between companies and their employees.
Source: Fast Company, April 2026
When people are quietly working around a company's AI strategy, leadership needs to understand why. Dismissing those actions as resistance risks overlooking the concerns that may be driving them.
Acknowledging concerns is a start. What changes afterward?
The HBR authors suggest three approaches: acknowledge employees' mixed feelings, create opportunities for experimentation, and reinvest the benefits AI produces.
All three are worthwhile. But there's a difference between agreeing with those ideas and putting them into practice.
A leader can hold a town hall, acknowledge that AI creates uncertainty, and encourage employees to embrace the change. That's a reasonable conversation to have, but what happens when employees return to their desks?
Do they know which tools they're expected to use? Do they understand what information is safe to share? Has anyone shown them how AI applies to their responsibilities? And do they know how success will be measured?
Without answers, employees are left trying to interpret what leadership wants. It's difficult to feel confident about a change when so much of it remains undefined.
Give people a chance to learn by doing
One recommendation from the HBR article that's particularly important is creating space for experimentation. But that space needs to be useful.
Most people have attended training sessions where someone demonstrates an impressive technology, answers a few questions, and sends everyone away with a link to documentation. People leave interested, but often struggle to apply what they've seen once they're back at work.
Atlassian's Teamwork Lab research found that formal training courses and self-service knowledge hubs were among the less effective approaches to developing human-AI collaboration, despite being widely offered. Its research favors hands-on experimentation and opportunities for teams to learn together.
Source: Atlassian, How to Make Space for AI Experimentation on Your Team
Someone in finance shouldn't have to spend hours figuring out how a generic AI demonstration applies to the reports they prepare every month. A marketing manager shouldn't have to translate examples from another industry into their own campaigns.
Bring the learning into their work. Sit with employees, understand what they're trying to accomplish, and help them experiment with AI using tasks they recognize. Give them the opportunity to question the results and see where the technology struggles.
People develop confidence when they have the chance to use a tool successfully, understand its limitations, and decide where it belongs in their work.