The AI-related leadership crisis that’s only five years away – and how to avoid it

There’s a talent problem hiding in plain sight. It doesn’t show up in this quarter’s earnings call. It won’t surface in next year’s workforce plan. But if organizations don’t start treating it like a design flaw now, they’ll run into a leadership wall within the next five years.

For all its power to make us more productive, AI is systematically reshaping the entry-level experience that traditionally trained the next generation of leaders.

We’re already seeing the market signals. Harvard University research indicates junior employment has fallen 9%, with entry-level hiring dipping 80% per quarter since 2023, at organizations adopting generative AI. ZipRecruiter’s 2026 Graduate Report found the share of entry-level jobs dropped to 38.6% at the start of 2026, down from over 44% three years ago.

Starter tasks are being automated or pushed upward, creating workload strain in the layers above that will become a clog in the leadership pipeline.

Entry-level roles have never been only about output. They are, in a very real sense, structured learning environments.

The junior analyst who builds the model manually develops an intuition for when the model is wrong. The new hire who drafts the memo and gets it marked up learns judgement by seeing their thinking challenged. The new college graduate who handles the messy client situation builds emotional resilience that no training course can replicate.

Yet addressing AI’s impact on early careers, the World Economic Forum paints a picture of a generation now entering careers where structured, repetitive tasks that once built confidence and competence are being automated away. From day one, young professionals now are expected to contribute analysis, insight and adaptability, often without the scaffolding that makes it possible.

Cornerstone’s recent survey of 2,000 workers found Gen Z reporting the highest rates of AI-driven role transformation, with 38% saying AI has fundamentally changed what their job requires. Yet they’re least likely to have received formal training to navigate that change. Most Gen Z workers using AI at work (59%) say their organization has never provided formal training for it.

This runs directly counter to the assumption that younger workers are better served by their organizations on AI. They’re forced to be self-reliant, and that’s where shadow AI becomes a risk multiplier, as employees use powerful tools outside approved guardrails because official enablement hasn’t kept pace.

What’s disappearing isn’t just work. It’s practice. Entry-level roles traditionally provided structured reps: learning how to prioritize, recover from mistakes, communicate uncertainty and build credibility. When AI absorbs the repetitive work, organizations may see efficiency but lose the developmental pathway that produces competent managers and grounded leaders.

AI is now table-stakes operating infrastructure. A few simple steps can ensure leadership development is designed, measured and continuously improved to become its foundation:

When AI takes tasks, don’t just delete them. Convert them into judgement loops that review outputs, validate assumptions, pressure-test recommendations and escalate edge cases. AI can be amazing or disastrously wrong, and someone has to build the muscle to tell the difference.

Use AI to generate drafts, options and scenarios. But require humans (especially fresh talent) to critique, refine and justify decisions. The learning is in the evaluation.

The market is shifting toward giving entry-level hires bigger responsibilities earlier. That can accelerate growth or create avoidable failure.

Build a progression ranging from low stakes and supervised decisions to independent ownership, with clear standards for what “good” looks like at each stage. AI agents can coach in the flow of work, suggesting resources, examples and next-step guidance at the moment of need, so development doesn’t rely on formal training employees rarely have time to take.

If early-career work is changing, managers must evolve with it. They need to explicitly teach how to think, not just what to do. That includes creating norms for responsible AI use, when to escalate, and how to communicate uncertainty.

Continuous workforce intelligence generated by AI models also can help managers see capability growth (or gaps) earlier, and trigger timely coaching, projects or learning before performance issues become attrition.

In fast-changing environments, a vehicle for building leaders is a stretch assignment like a rotation or gig. These cross-functional projects build judgement skills with practice; they create broader context and judgement faster than static roles do. Within our organization, for example, our Cornerstone Gigs programme creates short-term opportunities for employees to apply to assignments beyond their core team. It helps people stretch into new work that expands skills and builds new ones, while giving the business fresh perspectives and access to capabilities it might not otherwise surface.

AI can surface internal opportunities and match people to them based on evolving skills, so mobility becomes a system, not an informal spreadsheet or favour network.

Young professionals need permission and boundaries that disclose AI assistance where appropriate, validate outputs and discourage the outsourcing of thought. If every idea gets filtered through AI, people risk never developing their own instincts.

Embed lightweight governance like approved tools, safe-use prompts and checklists to guide good practice.

The real competitive divide won’t be how AI is adopted, but whether organizations build the human development engine that keeps pace. Automation without human judgement doesn’t scale performance; it scales mistakes.

Think of this shift as moving from doing the homework to grading it. Entry-level workers in an AI-augmented environment are increasingly positioned as evaluators and decision-makers, not just producers. Grading requires a standard, enough domain knowledge and critical thinking to catch when something feels off, and confidence to override a system that sounds authoritative but isn't.

The World Economic Forum’s Future of Jobs research finds that 39% of core skills will change by 2030. As AI rewires work, it also revamps capability. Organizations that replace lost entry-level reps with deliberate judgement-building will create leadership capacity that compounds with their technology.

The rest may scale efficiency today – but discover the leadership bill comes due tomorrow.