As artificial intelligence becomes embedded in everyday business decisions, a familiar reassurance is often offered: there will always be a human in the loop. But human oversight is not the same as human judgement.

Blueprint’s principle of subsidiarity is rooted in the idea of freedom with responsibility. It asks whether people remain able to exercise judgement, take ownership of decisions and be accountable for their consequences. Applied to AI, the question is simple: are we using technology to strengthen human agency, or replace it?

The risk of dependency

Recent research published in Harvard Business Review warns that poorly designed AI systems can encourage “cognitive offloading”, where people increasingly accept AI-generated outputs without question. The authors argue that organisations should design AI systems that strengthen reasoning and critical thinking, rather than simply maximise efficiency.

The risk is not that AI makes all the wrong decisions. It is that people gradually stop exercising the judgement needed to challenge them.

This may be an unintended consequence of a technology largely designed around speed, convenience and optimisation. The more friction we remove from decision-making, the less opportunity people have to develop the skills that good decision-making requires.

Why judgement matters

This is particularly important for early-career employees. For generations, professional expertise has been developed through practice: gathering evidence, testing assumptions, making mistakes and learning from them. Many of these formative experiences are now among the tasks most susceptible to automation.

Research conducted by KPMG and the University of Texas found that the strongest performers were not those with the highest levels of AI literacy. Instead, they were the individuals who actively questioned, refined and improved AI outputs. They treated AI as a tool rather than an authority. Those who simply accepted outputs with limited scrutiny often added little value beyond what the AI could produce alone. The capability that mattered most was not technical expertise. It was judgement.

For organisations, this raises an uncomfortable question: if AI increasingly performs the work through which people traditionally develop expertise, where will the next generation learn to exercise good judgement?

What kind of people are we becoming?

The debate about AI often centres on what the technology can do. Subsidiarity invites us to ask a different question: what is it doing to us? Every technology shapes behaviour. Over time, it influences what skills we practise, which responsibilities we retain, and how we relate to one another. The question is not simply whether AI can make better decisions. It is whether people will continue to have opportunities to develop the judgement needed to make decisions for themselves.

Subsidiarity begins with a belief that people grow through exercising responsibility. Judgement is not something we possess automatically. It is something we develop through practice. As AI becomes more capable, organisations will face choices about which decisions remain with people, where accountability sits, and whether individuals continue to have meaningful opportunities to exercise discretion and learn from experience.

 

For more see our Framework for human-centred AI – A guide for business.

The other blogs in our series van be read here