AI is changing how we work. But are we forgetting how we learn?

PwC’s latest investment in critical thinking raises a bigger question for leaders: as AI takes on more of the work, how do we protect and develop the human capabilities that make that work valuable?

I’ve seen quite a bit of discussion in the media and on social about PwC’s recent announcement, November 9 2026, that it is developing the critical thinking and judgement skills of over 85,000 employees*¹ at an allegedly investment of more than $100 million.

And – not only is it clear from the volume of debate – but I think they’ve touched on something really important.

Not because we’re suddenly discovering that critical thinking matters. But because AI is fundamentally changing how people develop it.

Think back to the early years of your career (unless like me, you realise that stretches further back than you care to realise). The time you took to research your topic and pull it together. The analysis and hypotheses you took – which often led you down a rabbit hole. The first draft (and often multiple reviewers) that came back covered in feedback. The decisions you made, questioned, and occasionally wished you’d made differently.

Not always enjoyable at the time, granted (!), but those experiences were formative. They helped us develop judgement, understand context, challenge assumptions and learn when something didn’t quite add up. As well as how to manage debate, corporate relationships and stakeholders – knowing when to challenge and when to take the feedback.

So, if AI increasingly does that groundwork for us, where does the learning happen?

The efficiency paradox

We’re rightly excited about what AI can do for productivity. It can accelerate large volumes of research, draft content, analyse information and take on work that previously consumed hours. And it’s already becoming part of everyday life – inside and outside of work.

Research from Gallup², found that among US employees using AI at work, the most common workplace uses, included writing and editing (51%) and for search and research (49%).

But here’s the paradox. Some of the very activities we’re automating are also the activities through which we’ve traditionally developed expertise.

That’s not an argument for making people do things the hard way just for the sake of it. Far from it, but it does raise an uncomfortable question: Could our pursuit of efficiency inadvertently remove some of the experiences that help people become better thinkers, decision-makers and leaders?

Human capability doesn’t develop by accident

This is where I think PwC’s approach is smart. Rather than simply teaching employees how to use AI more effectively, they’re investing in their ability to question its outputs, challenge conclusions and exercise independent judgement.

Importantly, the programme includes practical simulations and opportunities to apply those skills in everyday work, because knowing about critical thinking isn’t the same as actually doing it.

Something we’ve been exploring at Forty1 is how behavioural science can help organisations turn learning and good intentions into meaningful, sustained behaviour change.

Our Principal Behavioural Scientist, Dr Jess Holt, recently highlighted some of the psychological factors influencing AI adoption, from trust in the technology to concerns about professional identity and the value people bring.

In fact, research published in 2025³ found that people anticipated being perceived as less competent and more replaceable when using AI rather than conventional workplace tools.

That’s a fascinating tension, right? We’re encouraging employees to embrace AI, yet some may worry that using it undermines the very expertise they’ve spent years developing. Simply providing more training won’t necessarily resolve that.

We need to help people understand the value of their uniquely human contribution, build confidence through experience and create environments where questioning, experimentation and constructive challenge are encouraged. And crucially, where those behaviours are reinforced by leaders.

From teaching skills to creating the conditions to use them

During Human Advantage Week, Forty1 Founder, Richard Burton spoke with Brian Murphy, Global Head of Learning, Leadership & Talent Development at NTT DATA⁴, about the human capabilities that will matter most as AI reshapes the workplace.

It’s a conversation that feels particularly relevant to PwC’s announcement, because this isn’t simply about adding more learning modules to an already crowded development curriculum. It’s about rethinking how people learn through work itself.

That might mean giving junior colleagues opportunities to challenge AI-generated analysis rather than simply accept it. Creating space for teams to debate decisions and explain their reasoning. Encouraging leaders to share mistakes and demonstrate how they arrive at conclusions. And recognising not just the speed of someone’s output, but the quality of their thinking.

All of which requires something beyond individual capability: an organisational culture that genuinely values independent judgement.

The real Human Advantage

At Forty1, we talk about Human Advantage: the belief that in a world where intelligence is increasingly available on tap, lasting competitive advantage will come from the mindsets, commitment, skills and culture of the people using it. PwC’s announcement is an interesting illustration of that principle, but I think the challenge goes further.

If AI is taking away the work through which we traditionally developed judgement, we need to be far more intentional about how we build it. The goal shouldn’t simply be to create a workforce that’s better at using AI. It should be to create a workforce that’s better at thinking, questioning, connecting and deciding with AI.

Perhaps the question every leadership team should be asking is: Are we redesigning how our people develop as quickly as we’re redesigning how they work?

For more perspectives on developing human capabilities in an AI-powered world, listen to Richard Burton’s conversation with Brian Murphy as part of Forty1’s Human Advantage podcast series.

References:

¹PwC (2026) — PwC’s Critical Thinking & Judgment Certificate announcement

²Kemp, A. Gallup (2026, April 12) — Organizational AI Adoption Jumps Six Points.

³Reif, J. A., Larrick, R. P. & Soll, J. B. (2025) — Evidence of a social evaluation penalty for using AI. Proceedings of the National Academy of Sciences, 122(19).

⁴Forty1 (2026) — Human Advantage podcast series (Brian Murphy)