Where Does Judgement Get Formed Now?

As AI changes entry-level work, young people need more than technical training. They need environments where reasoning, uncertainty and professional judgement can be practised.

The UK's new Life Sciences Jobs Plan sets out how the sector can build the workforce it needs: more people, better routes in, closer alignment between education and employers. It is a useful starting point, but reading it left us with a more specific question, one the Plan raises without quite naming: not how many people the system produces, but where those people learn to exercise judgement in the first place.

The Plan's own evidence includes a detail many employers will recognise: there is a persistent gap between what a newly graduated BSc student knows and what employers expect on day one. The Plan traces this first to hands-on laboratory skills, and also to communication and an understanding of the commercial environment in which the work sits. MSc graduates are generally seen as more work-ready. Employers are left choosing between an inexperienced hire who needs upskilling, or a more experienced candidate who may already have the relevant experience.

That gap has always existed to some degree.

What's changed is what used to close it.

For most of the last few decades, early-career roles did two things at once. They produced output, and they built the judgement of the person producing it. A graduate joined a team, made small decisions under supervision, got some of them wrong, watched an experienced colleague reason through something genuinely unresolved, and slowly built the pattern-recognition no course could give them directly. Nobody designed this. It was simply what happened when junior people did real work alongside senior people, repeatedly, in a live organisational context.

AI may be changing parts of that arrangement. Many of the tasks once assigned to juniors — drafting, first-pass analysis, summarising, routine problem-solving — can now be done faster by a senior person prompting a model than by delegating to someone learning the ropes. That is a genuine gain in productivity. It may also be altering the informal training ground that used to exist inside the job itself.

The Plan recognises part of this. It backs a new Early Careers Jobs Alliance, funded at £20 million, specifically to work out how AI is reshaping entry-level work and to develop practical approaches that protect and expand routes into employment. The fact that a government workforce plan includes this kind of enquiry is telling: the entry-level rung is changing faster than existing routes can easily adapt.

This is not simply a skills gap, and it is not solved by teaching AI tools alongside the existing curriculum. What employers describe is harder to define: the ability to interpret context, weigh competing considerations, recognise uncertainty, ask better questions, and make a sensible call when there is no obvious right answer. Those qualities develop through experience, but not through time alone. They depend on the environments people spend time in.

This is the problem CognateHelix is designed to respond to. Rather than treating AI literacy as a module to be taught, or work experience as a box to be ticked, CognateHelix is built around repeated engagement with real, unresolved organisational questions that bring together technical, organisational and human considerations. In an early pilot, Outbreak, students worked through an AI-assisted decision-making scenario with expert input and were asked to explain how they reached a conclusion, not simply what the conclusion was. The aim is not a polished output. It is to make a young person's thinking visible, to themselves and to the people who will eventually need to trust it.

The model is still being tested. We are working with a small number of schools and employers to understand which forms of engagement genuinely build judgement, and which merely simulate it. Early signals are encouraging, but we are treating them as indicative, not conclusive, and we expect the design to keep changing as we learn more.

It is where, now, they will have the chance to practise thinking for themselves, before the stakes are real.