Editorial
When a group of employees at a technology consultancy in China were given access to ChatGPT for a week, something curious happened. Their supervisors, along with external evaluators, judged their work to be significantly more creative, more novel and more useful than that of their peers without access to AI.
The results, published in the Journal of Applied Psychology, came with a revealing footnote: the gains only materialized for employees with high metacognitive skill. In other words, AI did not make everyone smarter; it made self-aware thinkers more effective.
This small but rigorously designed field experiment reflects a broader truth emerging in today’s AI-enabled workplace: tools matter, but how they are used matters more. As companies invest billions in large language models (LLMs) to boost productivity and innovation, a quiet variable may separate those who thrive from those who merely automate: the ability to think about one’s thinking.
CHROs and Chief Learning Officers, take note: If AI is the engine, metacognition is the gear shift. It enables knowledge workers not only to complete tasks, but to adapt, reflect and iterate, i.e., to learn how to learn in a machine-augmented world.
GenAI Expands the Mind (if the Mind Is Ready)
Metacognition, a term drawn from cognitive psychology, refers to the ability to plan, monitor and evaluate one’s own thinking. It is what allows a designer to recognize when a creative block is self-imposed, or a product manager to question whether the metrics they are tracking truly reflect user experience.
According to Anne-Laure Le Cunff, a neuroscientist and founder of Ness Labs, metacognition functions as a “Swiss Army knife for the mind.” It operates through three intertwined elements:
- Metacognitive knowledge: Understanding one’s strengths, weaknesses and the tools available.
- Metacognitive regulation: The act of applying that knowledge through planning, reflection, and strategic adjustment.
- Metacognitive experience: The emotional awareness that accompanies learning and problem-solving.
In the Chinese field experiment, metacognition proved pivotal. Employees who lacked the ability to step back and evaluate their own cognitive approach saw little benefit from LLM access. Those with high metacognitive skill, however, reported greater cognitive job resources (including better information access, easier task-switching and more mental bandwidth) and produced measurably more creative work.
This is not surprising. GenAI thrives in environments rich in experimentation, feedback and adaptive learning. It augments what is already there, not by replacing cognition, but by enhancing the conditions under which it operates. A weak cognitive strategy plus AI yields faster mediocrity. A strong cognitive strategy plus AI creates exponential insight.
From Mindless Productivity to Mindful Exploration
Much of the current discourse around AI and productivity remains fixated on speed: shorter drafts, quicker code, more meetings turned into transcripts. But Le Cunff warns against what she calls “mindless productivity” — an obsession with output that disregards quality, context and mental cost. The more powerful AI becomes, the more dangerous this mindset gets.
By contrast, mindful productivity, grounded in metacognitive awareness, shifts attention to learning velocity and decision quality. It encourages workers to ask better questions, choose tools with discernment and pace their thinking in ways that preserve creativity and reduce burnout. Notably, Le Cunff’s framework aligns with the neuroscience of learning: the brain performs best when it cycles between focused attention and diffused reflection, i.e., deep work punctuated by wandering insight.
AI can support this rhythm, but only if work is redesigned to allow it. Managers who flood calendars with status meetings or reward visible busyness will find themselves automating shallow work. Those who create time for deep thinking and deliberate reflection will unleash far more.
To Achieve GenAI Gains, Design for Metacognition
The implications for human capital strategy are profound. Organizations seeking to harness AI for innovation must also build structures that support metacognitive behavior.
This begins with psychological safety. Many employees still use AI in secret, fearing that admitting reliance on a chatbot implies incompetence. Le Cunff suggests that organizations must actively normalize attribution, inviting teams to disclose how AI was used in producing deliverables. Such transparency not only reduces stigma but spreads effective practices.
Second, job design must shift from task completion to cognitive resource management. The most creative teams are not necessarily those with the most hours worked but those with the clearest oscillation between production and reflection. Time should be allocated not just for delivery, but for feedback, iteration and learning.
Third, L&D functions must train people not merely on how to use AI, but on how to think with it. This means embedding metacognitive prompts into daily workflows:
- What is the actual problem I’m trying to solve?
- What assumptions am I making?
- What would success look like, and how might I know if I’m wrong?
Even short interventions (journaling, post-project retrospectives and “Plus-Minus-Next” reviews) can raise metacognitive awareness. The results compound over time.
Metrics to Track Metacognition Progress
Metacognition may sound abstract, but it is measurable. Organizations can begin by tracking:
- Experiment Velocity: Number of AI-enabled experiments per team or quarter.
- Lesson Reuse: Frequency with which lessons from one project inform another.
- Psychological Safety: Employee agreement with statements like “I feel safe sharing how I use AI at work.”
- Creativity Outcomes: Manager and peer ratings of novelty and usefulness.
These are not vanity metrics. They reflect an organization’s ability to learn faster than the pace of external change.
The Strategic Case for Curiosity
In an era defined by disruption, the illusion of certainty becomes dangerous. Human brains crave clarity, but when environments change faster than mental models, certainty often signals delusion, not mastery.
Metacognition acts as a counterweight. It slows impulsive decision-making, highlights blind spots and encourages a mindset of curiosity. In Le Cunff’s words, it helps people approach uncertainty like scientists by asking better questions instead of rushing toward premature answers.
That mindset has organizational consequences. Companies that foster experimentation, normalize failure and train employees to reflect on how they think will be better positioned to harness AI’s power without succumbing to its risks. Those that don’t will likely find themselves overwhelmed, not by the tools themselves, but by the speed at which they’re misapplied.
Think Like a Machine, Reflect Like a Human
The proliferation of AI has prompted fears of cognitive obsolescence: Will humans still matter if machines can think? The better question is: What kind of thinking will remain uniquely human?
Conference
Sep
16
HR Tech Conference Las Vegas 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
The answer lies in metacognition. While machines can generate answers, they do not ask why a question matters. They do not reflect on their own reasoning or change course based on discomfort or doubt. They do not wonder what else might be true.
Humans do. And that capacity for reflective, recursive, self-aware thinking is not just a philosophical curiosity. It is the foundation of ethical decision-making, responsible leadership and long-term adaptability.
The Future of Learning Isn’t About Mastering AI. It’s About Mastering Ourselves
As GenAI becomes embedded in the fabric of work, the organizations that pull ahead will not be those with the largest models or fastest outputs. They will be the ones whose people can think with intention, adapt with grace and learn faster than change itself.
Metacognition gives workers the tools not only to improve their performance, but to understand the systems they operate in. It cultivates resilience amid complexity, discernment amid abundance, and perspective amid noise.
AI may accelerate what we do, but metacognition determines who we become. And to unlock our full potential, we must train the human, not just the model.
Editor's Note: Read more about how to balance human skills and AI capabilities below:
- Human Collaboration Still Matters in the Age of Artificial Intelligence — Businesses must marry the possibilities offered by AI with true human engagement.
- The Human Agency Scale: The AI Strategy Framework Business Leaders Need — To automate or not is no longer the question. The question is: what kind of partnership between AI and people will help this work shine?
- Waking Up to Our Power: Digital + Human Capabilities for a Future Ready Workforce — Beyond technical know-how, the future-ready worker needs a new blend of human and digital skills — anchored in awareness.
Learn how you can join our contributor community.
