“One of the worst mistakes you can make in a company is ‘de-weirding’ AI by treating it like an IT product.”

Ethan Mollick is one of the foremost global experts in AI and an associate professor at The Wharton School. He doesn’t mince words when it comes to how HR leaders should be making decisions about the technology.

At UNLEASH America 2026 in Las Vegas, Mollick delivered a wide ranging keynote address, covering academic research and real-time examples of how to get high-value results from widely available AI tools.

Sitting down for an exclusive interview with UNLEASH, he was even more direct in his assertion that “AI is a General-Purpose Technology. It affects everything we do. It touches everything we do.”

The message for HR leaders is clear: AI needs to drive how work is fundamentally redesigned, rather than a narrow technology deployment.

One of the key themes from this year’s show was an admission from HR leaders that they are feeling overwhelmed by AI, due in part to how fast the technology is evolving.

When Mollick gave a keynote at UNLEASH America 2024, he called on HR leaders to “invite AI into everything you do.”

“When I last talked to you guys, we were in a phase I’d call ‘Co-Intelligence’, where you’d go back and forth with an AI chatbot or it’d help write documents,” Mollick explains.

“We’re now in a world where agentic systems are real. You can assign an AI to do seven or eight hours of work, and it will just do it for you. That’s a very different circumstance.”

What’s needed is a change in how work operates, but Mollick says he worries that, in times like these, HR leaders “focus on what’s familiar.”

At a time when some organizations are pressing ahead with genuine AI-based transformation, others are stuck “trying to get approval from Legal” or are blinkered by a myopic focus on productivity.

“It’s part of a general problem that is in a lot of HR leaders’ heads right now, but part of that is about redesigning process and work to bring out purpose,” he explains.

I worry this is viewed as reactive change management.”

Eating risk and developing purpose – the foundations of AI success

Ethan Mollick delivers his keynote address at UNLEASH America 2026

What’s different now from previous challenges facing HR leaders is that they cannot simply go to market and purchase the technology solution that fixes the problem.

As Mollick puts it: “There’s no special tool out there; you have to invent while you’re there. You can’t just react.”

Which means HR leaders and practitioners need to be willing to eat risk to avoid turning the function into another risk-reduction exercise, similar to that performed by Finance or Legal.

Being willing and able to accept risky innovation and experimentation is a huge deal for HR leaders.

Crucial to this is developing second-order thinking – a framework that analyzes the long-term consequences of actions, not just the immediate results.

Closing out his keynote address at UNLEASH America this year, Mollick urged HR leaders to share positive examples of AI integration and to empower their teams to leverage AI effectively.

Such actions can inspire purpose – the absence of which has become the fastest-growing employee complaint in the past few years, coinciding with greater use of AI.

“People talk about purpose in a very surface way, but purpose is also about this transformation piece,” Mollick says.

He tells UNLEASH that HR is often caught up in worrying about bottlenecks concerning purpose in the future of work, highlighting programming as an example.

“If you talk to coders today, the best coders aren’t coding anymore. And that’s happened in something like the last two months, it’s not a long-standing thing,” he details.

“You’re in a world where the best coders aren’t coding anymore, and now they have to figure out the reskilling within their own job.”

“There’s a second-order capability of changing with change that I’m worried is missing a lot of in these cases.”

For HR leaders right now, that means acknowledging that the AI of today will not be the same tomorrow and adapting accordingly – in Mollick’s words, that is “not where we’re headed.”

“We are on an exponential curve, so you have to build a second-order ability to change as change happens,” he explains.

It’s a teammate already. If you have not figured out a way in your organization to use AI to actually delegate work to AI, then you’re not experimenting hard enough.”