Why are workers more productive with AI, but not organisations?
We need to have a serious talk about how we’re using AI. We know that the majority of workers it every day:
87% of people use AI at work. 75% say it makes them more productive
However:
Only 13% believe their organisation is performing better
What on earth is going on? There’s a growing body of research that claims to know the answer, and it has to do with how we use our brains when we work with artificial intelligence.
What is cognitive surrender when working with AI?
Most of us are familiar with the term ‘cognitive offloading’, where we use external tools to reduce the mental effort needed to complete a task.
Think of using Google Maps instead of memorising directions or setting calendar reminders instead of trying to remember every appointment. But cognitive surrender is different; it goes way beyond offloading.
A 2026 study from The Wharton School defines the term as: "Adopting AI outputs with minimal scrutiny, overriding intuition and deliberation.”
"Using a calculator offloads a specific math task while human reasoning stays in charge. Cognitive surrender is different: it is the moment when AI is not just doing a specialised task but making the decision, and the person adopts that decision as their own without recognising the transfer has occurred."
– Gideon Nave, PhD, Carlos and Rosa de la Cruz Associate Professor; Associate Professor of Marketing, The Wharton School
Why do we trust AI so easily?
Insights Organisational Psychologist Dr Tanya Boyd says, “When something looks polished, sounds confident and arrives quickly, as AI content often does, it's surprisingly easy to accept it without fully testing whether it's right.”
And she’s correct. We all have colleagues and friends who have practically deferred their entire personality to ChatGPT, presenting unverified AI stats to win every argument, often ignoring the reality that AI can hallucinate data and fails to consider context.
Why isn't AI delivering the productivity gains organisations expected?
It’s not a leap to hypothesise that the 62-point gap between workers feeling individually more productive and organisations plateauing at 13% could very well be caused by a high degree of cognitive surrender among workers using AI.
According to Glean’s Work AI report, there are two main ways people use AI, and both eat up productivity gains in different ways:
About engagement... It sounds like botsitters, people who take time to produce accurate and aligned output, should be the more engaged cohort, but the reverse is true.
For those motivated employees, there’s a “context tax”. Work AI states that, for every 10% more time workers spend feeding AI context, they are 25% more likely to feel “worn out”.
Add to that the very human reaction of not being thrilled to question or correct your cognitive-surrendering colleague’s work every day; it’s awkward and creates tension. It may lead to resentment when managers unknowingly praise work that appears high-quality but is actually poor.
No wonder 73% of botsitters are more likely to be looking for work.
But wait, can’t AI improve creativity?
It can indeed.
A 2024 study from the University of Exeter and UCL School of Management showed that, when used properly, generative AI enhances individual creativity (unfortunately, it still diminishes ‘collective novelty’).
The authors write:
"We find that access to generative AI ideas causes stories to be evaluated as more creative, better written, and more enjoyable, especially among less creative writers.”
But there’s a catch.
A follow-up 2025 MIT study breaks it down further, showing that the creativity gains are not automatic i.e. simply using ChatGPT won’t immediately make you a better writer or researcher. It’s all in how we use it.
"Employees who actively analyse their tasks, monitor their thinking, and adjust their approach, what is termed as metacognitive strategies, are significantly more likely to use generative AI in ways that foster creativity. This is because these individuals are better at using AI to access “cognitive job resources,” the raw materials that facilitate creative thinking.”
Why critical thinking matters more than prompting
"Metacognition — thinking about your thinking — is the missing link between simply using AI and using it well. It allows people to ask better questions, recognise knowledge gaps, and extract real value from AI tools instead of relying on them passively”
– Jackson Lu, General Motors Professor of Management at MIT Sloan
And, despite its importance, there’s a perceived lack of critical thinking in the workplace right now:
A 2026 internal survey of Insights clients found that 73% of respondents felt that it was ‘a top missing skill’ in their organisations, and over half (52%) ranked it as the top missing skill.
Again, thinking about that 62-point gap; it starts to make sense.
How can organisations reduce cognitive surrender and improve performance?
When talking about the MIT creativity study, Jackson Lu said,
"Even the most powerful AI won’t boost creativity if employees don’t know how to use it effectively. Organisations should pair AI deployment with metacognition training to maximise the creative benefits."
This is why Insights recently added a new application to Insights Discovery that highlights how our preferences shape the way we think and potential biases that may emerge in our decision-making process.
By understanding our critical thinking style, we can make more balanced decisions, work more effectively with others and improve our judgment over time.
Insights Discovery is a psychometric-based experience built on a four-colour model to describe and communicate different personality traits and preferences, basically how we communicate and work with others in different environments.
Each colour energy brings a combination of strengths and weaknesses to the table, and those attributes transfer to thinking patterns as well.
How Insights Discovery develops critical thinking
Insights Discovery Critical Thinking walks learners through a decision-making framework to help them make sound judgment calls, particularly around how they use AI output.
The programme builds on the principle that organisational performance doesn't come from producing answers alone; it comes from how well we evaluate those answers, build on them, challenge them, and turn them into decisions and actions that we’re willing to own.
Human contribution becomes more important because, as AI becomes more capable of generating content and -increasingly through agents- taking action on our behalf, humans show up later in the process; further away from the thinking behind what's been produced.
The question evolves from ‘can we generate an answer?’ into ‘can we distinguish a weak answer from a good one, are we willing to challenge a flawed assumption and can we decide effectively what to do next?’
Success depends on when you begin, and how far down the road of cognitive surrender and botshipping your employees have already travelled; the sooner the better.
But even for teams where cognitive surrender has become a way of working, this programme is a meaningful way forward that offers positive outcomes far and beyond better decisions. Things like higher engagement, more transparency, better collaboration and inspired innovation; all important measurements in the annual employee survey.
Dr Boyd, here at Insights, has a holistic perspective on the whole issue. She says:
Human skills like critical thinking and teamwork are strategic. They are how organisations turn productivity into performance. We know that, in order to build and activate these skills, we must start with self-awareness. Better thinking doesn’t just depend on the quality of the information in front of us, it also depends on our ability to understand ourselves and work effectively with the people around us”
Resources used in this article:
Kosmyna, N et al. Thinking Fast, Slow and Artificial: Generative AI Reduces Cognitive Effort While Thinking. SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6097646
Glean. Work AI Index. https://www.glean.com/work-ai-institute/reports/work-ai-index
Brynjolfsson, E, Rock, D and Syverson, C. The Productivity J-Curve: How Intangibles Complement General Purpose Technologies. American Economic Journal: Macroeconomics. https://www.aeaweb.org/articles?id=10.1257/mac.20180386
Davenport, T. Ten Reasons Why We Won't See Productivity. Substack. https://tdavenport.substack.com/p/ten-reasons-why-we-wont-see-productivity
Doshi, A R and Hauser, O P. Generative AI Enhances Individual Creativity but Reduces the Collective Diversity of Novel Content. Science Advances. https://www.science.org/doi/10.1126/sciadv.adn5290
MIT Sloan School of Management. Does generative AI actually enhance creativity in the workplace? https://mitsloan.mit.edu/press/does-generative-ai-actually-enhance-creativity-workplace
