AI promises speed, clarity, and scale, and in many ways, it delivers. But we don’t often think about what the price is for that efficiency. I don’t just mean the estimated $7 trillion dollars that will need to be spent on data centres in the next couple of years to keep up with demand.
The chainsaw effect
When the chainsaw was introduced, it transformed forestry. Work that once took hours could be done in minutes. It reduced physical effort, increased output, and changed the economics of the entire industry. However, it also changed the skills of the worker.
If you never learn to use a handsaw, you never build the same muscle memory, judgement, or feel for the material. You become dependent on the chainsaw. When it works, you’re fast, but when it fails, you either have no idea how to do the job, or you have to go out to get the right tools and learn how to.
AI is like a cognitive chainsaw. It allows us to cut through complexity quickly, brainstorm faster, draft faster, analyse faster and decide faster. What it’s also doing is reshaping the very capabilities we rely on when the stakes are high, the data is incomplete, or the tool gives us the wrong answer. And this is the cost.
Cognitive offloading
Humans have always offloaded our thinking. We write things down, we use calculators. and we rely on systems. This is not new. What is new is the scale of this offloading.
Research on cognitive offloading shows that when we expect information to be available externally, we are less likely to encode and process it. In other words, when the answer is just a few prompts away, our brains don’t bother doing the heavy lifting – it knows it doesn’t need to. Our brains are nothing if not efficient, but that efficiency comes with a trade-off.
Cognitive effort is a mechanism for learning, pattern recognition, and judgement formation. Emerging research in neuroscience suggests that effortful engagement is linked to stronger dopaminergic reward responses than passive or automated completion. In plain language, that means that when we work through a problem, rather than bypass it, we experience greater intrinsic reward, which reinforces learning and motivation over time.
While AI makes thinking easier it also reduces the parts of our work that make us feel intellectually stimulated and fulfilled. Research has suggested that one of the mechanisms behind this might be agency, or our sense of control over our work, which is a principal driver of intrinsic motivation. When we don’t feel like we’re fully in charge of our work, we feel disconnected. One study from 2025 found that those who worked with AI and then moved to natural intelligence, or, as I’m shortening it, NI, had 11% less motivation and were 20% more bored. It’s like usually flying from Brisbane to Sydney, but then needing to drive because there are no flights available. It can feel like a drag.
Neuroplasticity: Use it or lose it
Our brains constantly adapt, depending on how they are being used. Neuroplasticity research consistently shows that repeated behaviours strengthen neural pathways, while unused ones weaken over time. This is how we build expertise and also how we lose it. A now well-known MIT study called ‘Your brain on ChatGPT’ found that frequent use of AI tools can result in what the authors called “skill atrophy” in critical tasks like brainstorming and problem solving.
If AI is increasingly doing the structuring, synthesising, and generating for us, we may start to find it difficult to do things like:
- formulate a problem from scratch,
- sit with incomplete or ambiguous information, and
- generate options without prompts.
If we don’t use these capabilities, we will lose them. That loss might be gradual and unconscious, but one day we will be in a high-stakes moment ,sans AI, and we’ll find that our cognitive ‘muscle’ isn’t as strong as it used to be.
The temptation of certainty
One of the most critical capabilities in our wacky world right now, especially in relation to decision-making, is ambiguity tolerance. It’s our ability to sit with uncertainty, conflicting information, and incomplete data without rushing to closure, and it’s a marker of strong judgement.
AI, by design, reduces ambiguity buy providing ideas, answers, structuring thinking and filling in gaps. This assistance can be very helpful, however it can also create a dependency because our muscle for doing it has atrophied.
If every question is quickly met with a well-structured response, we may become less comfortable in the messy middle, which is where many of the most important decisions live.
In organisational contexts, this is important, because when leaders lose tolerance for ambiguity, decisions tend to converge too quickly, rely on surface level logic, or miss second and third order implications. In other words, decisions become efficient, but not necessarily effective.
The hidden cost curve
The challenge with all of this is timing. The benefits of AI are immediate, however the costs are often delayed. It’s like putting everything on a credit card. At first, it feels easy, convenient, and even smart. Eventually, though, the bill arrives – sometimes, with interest.
I’m not suggesting anyone should reject AI – that boat sailed long ago. Rather, it’s a call to be deliberate and intentional about where you use it, and where you don’t.
Five ways to protect your thinking (without giving up efficiency)
- Do the first pass yourself
Before using AI, spend a few minutes framing the problem, generating options, or forming a view. This preserves your ability to think independently before being influenced.
- Create intentional friction
Not every task needs to be outsourced to AI. Choose specific tasks to apply ‘slow think’ to, especially for complex or strategic decisions.
- Map the cost across time horizons
For any AI use, ask:
- What am I gaining in the short term?
- What capability might I be outsourcing in the medium term?
- What could this outsourcing cost me in the long term?
- Practice ambiguity exposure
Deliberately sit with incomplete information before seeking answers. Build tolerance for not knowing, without rushing to resolution.
So what:
- AI is not just changing what we do, it’s changing how we think, how we learn, and how we decide. Like any powerful tool, its impact will not just be determined by what it can do, but by how consciously we choose to use it.
- Here’s a simple question to sit with this week: Where have I gained efficiency, but unintentionally given something up in return?
- If the bill hasn’t arrived yet, what might it look like when it does?
References:
Doidge, N. (2007). The brain that changes itself. Viking Press.
Inzlicht, M., Shenhav, A., & Olivola, C. Y. (2018). The effort paradox: Effort is both costly and valued. Trends in Cognitive Sciences, 22(4), 337–349.
Kosmyna, Nataliya & Hauptmann, Eugene & Yuan, Ye & Situ, Jessica & Liao, Xian-Hao & Beresnitzky, Ashly & Braunstein, Iris & Maes, Pattie. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. 10.48550/arXiv.2506.08872.
Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688.
https://hbr.org/2025/05/research-gen-ai-makes-people-more-productive-and-less-motivated