Thinking About Thinking: The Real Secret to AI Innovation
Reflections on the recent Harvard Business Review article, 'Why AI Boosts Creativity for Some Employees but Not Others'.
While the world seems obsessed with the latest iteration of software or the speed of a new processor, a recent Harvard Business Review study redirects our focus to the most powerful tool of all: the human mind. It explores a paradox that many of us have felt but perhaps could not quite name: why does generative AI make some employees brilliant while leaving others exactly where they started?
The research, which was recently published in the Journal of Applied Psychology, provides a definitive answer. The differentiator is not technical skill, nor is it the quality of the prompts being used. Instead, it is a psychological trait known as metacognition. This is, quite simply, the ability to think about your own thinking. It involves the active planning, monitoring, and evaluation of your cognitive processes as you work. For those interested in the future of work, this is the most important insight of the year.
A Deep Dive into the Research
The study, led by Shuhua Sun and Jackson G. Lu, was remarkably rigorous. It utilised a field experiment involving 250 employees at a technology consulting firm, followed by controlled experiments with over 1,000 participants. This methodology is particularly robust because it combines real-world observation with the precision of a laboratory setting.
The researchers discovered that generative AI only boosts creativity for individuals with high metacognition. These individuals use AI to acquire what the study calls cognitive job resources. Specifically, they use the technology to expand their knowledge base and free up mental capacity for more complex tasks.
By contrast, employees with low metacognition tend to engage in passive consumption. They treat the AI as an oracle rather than a partner, accepting the first answer they receive without scrutiny. For these individuals, the AI actually does nothing to enhance their creativity. It may help them work faster, but it does not help them work better. This finding challenges the common assumption that AI is a 'rising tide that lifts all boats'. In reality, it appears that AI may actually widen the gap between strategic thinkers and those who merely follow instructions.
Roles, Industries, and Personalities
Certain environments and temperaments are naturally more conducive to this metacognitive approach.
Industries and Roles
We see the greatest impact in knowledge-intensive industries such as software engineering, strategic consulting, and research and development. In these fields, the 'right' answer is rarely obvious. Success requires the synthesis of disparate ideas and the constant questioning of assumptions.
Roles that involve design thinking or complex problem-solving lend themselves to metacognition because they require a high degree of cognitive flexibility. If you are an architect using AI to generate structural options, your value lies in your ability to evaluate those options against human needs and aesthetic principles. You are not just 'using' the tool; you are managing a complex feedback loop.
Personality Types
The research suggests that individuals high in Conscientiousness and Openness to Experience are most likely to thrive. Conscientious people have the discipline to monitor their own work and check AI outputs for errors. Meanwhile, those who are open to experience are more willing to experiment with the strange, divergent ideas that AI often produces. Interestingly, a sense of authenticity is also a strong predictor: those who are true to their own values are more likely to use AI as a tool for self-expression rather than a shortcut to conformity.
Practical Applications: Actionable Advice
If we want to close the 'creativity gap' in our organisations, we must move beyond simply providing access to tools. In my book, Enhanced Leadership, I argue that the primary responsibility of a modern leader is to develop the 'internal technology' of their team. We must help people master their own cognitive processes before we ask them to master an external algorithm.
Here is how we can increase metacognitive capacity at three distinct levels:
For Individuals
Individual growth starts with a 'strategic mindset'. This involves moving from a reflexive approach to a reflective one.
The Reflection Loop: After receiving an AI response, ask yourself: 'What is the most obvious flaw in this answer?' and 'What perspective has been completely ignored?'.
Metacognitive Questioning: Adopt the habit of asking strategy-eliciting questions during a task. 'Is there a way I could do this even better?' or 'What can I do to help myself right now?'.
The Monitoring Pause: Set a timer for twenty minutes. Every time it rings, stop for sixty seconds to evaluate your progress. Are you getting closer to a creative breakthrough, or are you just refining a mediocre idea?
Emerald Framework Application: Focus on metacognitive knowledge. This means understanding your own strengths and the specific demands of the task. Do not just start prompting; plan your strategy first.
For Oversight and Management
Oversight must shift from checking the final output to evaluating the 'human-AI joint activity'.
Prompt Transparency: Require team members to share not just their final work, but the sequence of prompts that led to it. This allows managers to identify where the thinking process might have stalled.
Peer Review of Thinking: In project reviews, ask team members to explain how they validated the AI’s suggestions. What did they reject, and why?
Cognitive Load Audits: Managers should monitor for cognitive offloading. If an employee is consistently producing 'generic' work, they may be relying too heavily on the AI. Re-engage their metacognition by assigning them a task that requires a high degree of personal synthesis.
For Organisational Culture
Culture is the most powerful lever for sustained innovation. We must build a 'thinking culture' that values the process as much as the product.
Incentivise Learning Velocity: Instead of rewarding speed, reward those who can show how they used AI to learn a new concept or solve a problem in a novel way.
Create a Shared Language: Adopt terms like 'metacognitive sensitivity' and 'cognitive resources' across the organisation. When everyone understands what 'thinking about thinking' means, it becomes a collective goal.
Psychological Safety for Dissent: Encourage a culture where it is safe to challenge 'the computer'. AI can often produce biased or shallow results; employees must feel empowered to call these out without fear of being seen as 'unproductive'.
Metacognition Training: Invest in training programmes that focus on the science of learning. Teach your employees about metacognitive experiences—those feelings of ease or difficulty that guide our decision-making.
Conclusion: The Future belongs to the Reflective
The HBR report is a vital reminder that technology is a multiplier, not a substitute.
Here is a link to the full HBR report: https://hbr.org/2026/01/why-ai-boosts-creativity-for-some-employees-but-not-others
If you multiply a passive mindset by the power of AI, you still have a passive result. But if you multiply a metacognitive, strategic mindset by that same power, the results are truly extraordinary.
The future of leadership is not about being the smartest person in the room; it is about being the most aware. I encourage you to explore these themes further on our website at levelupleadership.uk, where you can find our latest podcast episodes and more resources on cognitive excellence. Let's stop focusing on what the machines can do and start focusing on what we can do with them.
Follow on LinkedIn - Spotify - YouTube - Apple
Level Up Leadership is a passion project in my spare time. I enjoy doing it, and I intend to keep these articles and podcasts free. However, the software and equipment I use isn’t free! So, if you are enjoying this content and would like to make a donation, you can do so by clicking this button. Thank you.