If AI does the work people used to learn from, where does professional judgement come from?

Our September 8 Organisational Psychology Discussion Group was facilitated by Abby White and Paul Hulston from Deloitte, drawing on its 2026 Human Capital Trends research, its Gen Z and Millennial Survey, and client examples. A key conclusion was expertise isn’t built by knowing things. It’s built by doing things, getting feedback and reflecting on them.

The routine tasks AI is now absorbing, such as drafting, summarising, analysing and reviewing, were never just “admin”. They were where early-career professionals:

• practised in low-stakes conditions
• made small mistakes and got corrected
• built pattern recognition and tacit knowledge
• developed confidence in their own judgement

Remove those tasks without replacing the learning, and we risk a generation who can use the output but struggle to evaluate it. Evaluating AI output is the skill we most need people to have.

What psychology tells us about the way forward:

• Deliberate practice beats passive exposure. Experience alone doesn’t build expertise. Challenging tasks with timely feedback do. If AI removes the practice, organisations will need to build it back in on purpose, through simulations, case work and structured stretch assignments.
• Judgement is built in the struggle. Learning research suggests that effortful retrieval and problem-solving embed knowledge more durably than being handed answers. Over-relying on AI can short-circuit that process.
• Supervision is the original apprenticeship model. In psychology, no one becomes a practitioner by reading about it. We learn through supervised practice, reflection and feedback from people who model good judgement. Other professions may need to rethink this model for the AI era: mentoring, coaching and reflective practice treated as core infrastructure, not optional extras.
• Confidence and identity matter too. Capability depends on self-efficacy and professional identity as well as skill. Early-career people need chances to experience mastery, not just to supervise machines.

The question is, where are people in your organisation still getting the chance to practise, receive feedback and build judgement, and where have we quietly removed it?

The productivity gains from AI are real. But if we don’t design the apprenticeship of the future deliberately, we’ll find out in ten years that we’ve run short of leaders.

How is your organisation approaching this? 👇

#FutureOfWork #AI #OrganisationalPsychology #HumanCapital #EarlyCareers #LearningAndDevelopment #Leadership #Deloitte

Video