Need to write an email? Ask AI.
Summarise a report? Ask AI.
Analyse information, brainstorm ideas or make a recommendation? AI can help with those too.
Tasks that once required considerable effort can now be completed in minutes. In many ways, this is a positive development.
But as AI becomes increasingly capable, an uncomfortable question emerges:
Are we using AI to become better thinkers, or are we becoming comfortable letting it think for us?
The concern is not simply that AI might replace certain tasks. It is that we may become less accustomed to questioning, analysing and forming our own conclusions.
When Convenience Begins to Replace Critical Thinking
Technology has always helped us reduce effort. Calculators handle calculations, search engines retrieve information and digital tools simplify complex processes.
Generative AI introduces a different dynamic. It doesn't just retrieve information. It produces arguments, interpretations, recommendations and seemingly well-reasoned conclusions.
This creates opportunities to work efficiently, but also the temptation to accept answers without understanding how they were reached.
A 2025 study by researchers from Microsoft Research and Carnegie Mellon University examined generative AI and critical thinking among 319 knowledge workers, drawing on 936 examples of AI-assisted work.
The researchers found that greater confidence in AI was associated with less critical-thinking effort, while greater confidence in one's own abilities was associated with more critical thinking.
These findings do not prove that AI permanently weakens our thinking abilities. They do, however, raise questions about how reliance on AI changes our engagement with information.
When answers become easier to obtain, the ability to question them becomes even more important.
Getting the Right Answer Doesn't Mean We Understand It
Consider an employee preparing a recommendation for management.
They ask AI to analyse a situation, identify possible solutions and suggest the best course of action.
Within seconds, they receive a structured response with supporting arguments and a convincing conclusion.
But what happens when someone asks why that recommendation is appropriate? Can the employee explain its assumptions, identify missing information or recognise circumstances where it might not work? Or are they presenting an answer they did not fully evaluate?
A 2026 MIT study involving 67 participants examined whether AI-assisted discussions could improve people's ability to identify misinformation. While AI assistance helped participants assess information during the task, the benefits did not translate into lasting improvements in independent discernment.
The findings highlight the distinction between performing a task successfully with AI and developing the capability to perform it independently.
If you cannot explain or challenge an AI-generated recommendation, how confidently can you stand behind it?
What Happens When We Stop Practising?
For organisations, the implications extend beyond individual productivity.
As AI becomes integrated into workflows, employees may have fewer opportunities to practise skills that previously formed part of their responsibilities.
A junior employee relying on AI to analyse every problem may have fewer opportunities to develop independent reasoning. A manager consistently accepting AI-generated recommendations may become less accustomed to challenging assumptions. A team generating ideas with AI may still struggle to evaluate which are appropriate for its circumstances.
In its 2026 article, When Everyone Uses AI, Companies Risk Losing Critical Skills, Boston Consulting Group highlighted the potential risk of AI-related deskilling.
In a survey of 70 senior executives, half reported already observing deskilling within their organisations, while more than 60% considered it a material threat over the next three to five years.
The capabilities identified as particularly vulnerable included judgement, decision-making, problem framing and creative thinking.
Although these findings reflect executive perceptions rather than direct measurements of skill decline, they point to an important organisational challenge.
Efficiency matters, but so does protecting opportunities for employees to develop the capabilities AI increasingly performs.
AI Should Support Judgement, Not Replace It
The solution is not to discourage AI adoption.
AI can help individuals explore unfamiliar topics, challenge assumptions, consider alternative perspectives and work through complex information.
The difference lies in how it is used.
In their 2026 Harvard Business Review article, Don't Let AI Destroy the Skills That Make Your Company Competitive, Graham Kenny and Ganna Pogrebna argue that organisations risk weakening important capabilities when AI replaces opportunities for human judgement, deliberation and shared learning. Their central argument is that AI should strengthen human expertise rather than erode it.
Instead of asking AI to make every recommendation, we can use it to test our reasoning. Instead of accepting the first response, we can examine alternative interpretations. Instead of treating AI-generated information as a final answer, we can use it as a starting point for deeper inquiry.
For example, rather than immediately asking, "What should I do?", we might develop our own perspective and then ask, "What assumptions am I overlooking?"
This changes our role from passive recipients of information to active participants in the thinking process.
Before asking AI for a solution, form your own perspective. Then use it to challenge your assumptions and identify what you may have overlooked.
Building a Workplace That Still Values Independent Thinking
As organisations accelerate AI adoption, the focus cannot remain solely on how much faster employees complete tasks.
It must also consider whether employees continue developing the capabilities needed to work effectively without unquestioningly relying on technology.
This means creating opportunities to question AI-generated outputs, practise independent problem-solving, engage in meaningful discussions and exercise human judgement.
These practices are not about slowing progress. They ensure efficiency does not come at the expense of capability.
Microsoft Research has explored this direction through its work on AI as a tool for thought, examining how AI systems can support reflection and active reasoning rather than simply generating completed outputs.
The opportunity is not merely to help employees accomplish more. It is to help them become more capable in the process.
Is your organisation measuring only how efficiently employees use AI, or also whether their judgement and problem-solving capabilities are growing?
The Future of Work Still Requires Us to Think
AI will continue to reshape how we work, taking on tasks that once required substantial human effort and creating new opportunities.
But greater technological capability does not automatically translate into greater human capability.
Critical thinking, judgement and the ability to recognise complexity remain essential, particularly when answers are not straightforward.
For individuals, this means remaining actively involved in the reasoning behind their work. For organisations, it means looking beyond AI adoption and productivity metrics to consider how human capabilities are developed and sustained.
The real risk is not that AI becomes better at thinking. It is that we become less intentional about doing it ourselves.
Perhaps the question we should ask is no longer just, "How much work can AI do for us?"
But rather: "As AI becomes more capable, are we making sure our people do too?"