As AI reshapes the conditions of professional work, a quieter question is emerging: not whether expertise will be needed, but whether the conditions that form it still exist.

Recent reporting from Accenture, Jisc, and the Institute of Student Employers points to a subtle but consequential shift in early-career development. While large-scale entry-level job loss has not yet materialised, a quieter change may already be underway: growing uncertainty about how professional capability actually develops in AI-enabled environments.

The signal is not one of sudden disruption. It is one of weakening confidence in the developmental pathway itself.

The Erosion of the Traditional Training Ground

Historically, professional capability formed through participation, not simply instruction.

Many early-career tasks were routine and repetitive. Yet these environments provided something essential: structured exposure to judgement, correction, accountability, and gradually increasing responsibility. Over time, practitioners learned not only what to do, but how to interpret situations, recognise patterns, navigate ambiguity, and make decisions in context.

As AI systems increasingly automate or compress many routine functions, these developmental environments are beginning to change structurally. Several tensions are emerging: the space between highly assisted work and high-complexity responsibility is narrowing; tasks that once enabled gradual capability development are disappearing; and organisations risk losing the pipeline of developing expertise on which future capability depends.

This matters because expertise is not formed solely through knowledge acquisition. It develops through repeated engagement with uncertainty, consequence, interpretation, and reflective adjustment over time.

Learning Processes Under Pressure

The integration of AI into education and training introduces a further tension.

AI-supported feedback offers clear advantages: accessibility, speed, responsiveness, and lower barriers to iteration. But feedback is intended to support thinking — not replace it. When feedback becomes immediate, generative, and continuously available, learners may begin refining outputs before fully forming their own reasoning. In these environments, performance may improve while the underlying development of capability becomes harder to observe.

A learner may produce sophisticated work while relying heavily on external cognitive support that they cannot yet independently reproduce under conditions of uncertainty or incomplete information. This distinction becomes increasingly important in professional environments where AI-assisted performance is easy to generate, but responsibility for decisions remains human.

“Performance in an AI-enabled environment can be optimised. Capability — the kind that holds under pressure — still has to be formed.”

From Exposure to Intentional Design

For much of modern professional history, capability formation was embedded implicitly within organisational structures. People learned because the structure of work required them to.

If those structures are changing, capability may no longer emerge reliably as a byproduct of participation alone. This points towards more intentional environment design.

Future learning and early-career systems may need to place greater emphasis on:

• Protected space where independent reasoning can develop before AI optimisation dominates the process
• Opportunities for safe failure and reflection
• Gradual progression of responsibility
• Making reasoning processes visible — not only outputs

The challenge is not to remove AI from learning environments. AI is likely to become an enduring component of both education and work. The challenge is ensuring that human capability continues to form alongside increasingly capable systems.

Distinguishing Performance from Capability

As AI becomes more integrated into professional and educational workflows, the distinction between performance and capability becomes increasingly important.

Performance in AI-enabled environments tends to emphasise speed, fluency, and synthesised output — often optimised, consistent, and data-driven. Capability, in the human sense, involves something different: problem framing, contextual judgement, navigation of uncertainty, and the capacity to take responsibility for decisions where variables are hidden or incomplete.

This distinction is central to how CognateUK thinks about human development. Drawing on Professor Rose Luckin’s framework of the 7 Elements of Human Capability — from metacognitive intelligence to social and contextual adaptability — genuine capability is not a single skill. It is a layered system of capacities that develops through experience, reflection, and increasing responsibility over time. AI can support many of those elements. It cannot substitute for the formation process itself.

If organisations optimise primarily for visible performance outputs, there is a risk that expertise becomes increasingly brittle — outputs may appear highly capable while the underlying reasoning remains comparatively fragile.

A Question of Formation

The central challenge is no longer only how expertise is recognised. It is how expertise is formed.

Many of the developmental structures that historically supported professional growth emerged naturally through organisational participation, accumulated practice, and gradually extending responsibility. As AI reshapes those environments, the assumptions underpinning capability formation may also need to be reconsidered.

The question is not whether AI will change work. It already is. The deeper question is whether organisations, educators, and institutions can design environments in which judgement, reasoning, responsibility, and reflective capability continue to develop robustly alongside increasingly intelligent systems.

That challenge may prove less technological than developmental.

If these questions are relevant to your organisation or institution, we’d welcome the conversation. Explore our thinking on human capability at cognateuk.com or connect with us on LinkedIn.

This article reflects the working perspectives of CognateUK and is intended to support informed discussion. It does not constitute advice or represent the official positions of any affiliated organisations or partners.