AI can now summarise an annual report, build a financial model, compare competitors, and produce a polished investment thesis in minutes. For ordinary investors, this is genuinely useful. But it also raises an uncomfortable question: if a machine gives you the answer, are you becoming a better investor—or simply becoming better at accepting answers?
The issue is not technology itself. It is what happens when we repeatedly delegate the thinking part of investing to a tool.
The GPS lesson
Think about how people used to navigate an unfamiliar city. They noticed landmarks, remembered turns, checked direction, and slowly built a mental map. GPS makes this easier, but it also removes much of that mental work. You may reach the destination efficiently, yet never learn the route.
Melanie Challenger calls this broader phenomenon “deskilling”: when technology repeatedly performs a task for us, we may lose the regular practice that once kept that ability alive. Research on GPS use suggests an association between heavier reliance on navigation tools and a steeper decline in spatial memory over time. The evidence is not decisive, but the underlying idea deserves attention: convenience can quietly remove practice.
Investing has its own version of the blue line on a map.
The convenience trap
Earlier investors had to read annual reports, study competitors, calculate ratios, build models, and sit with uncertainty. Much of it was slow and tedious. But the slowness forced questions:
- How exactly does this company make money?
- Is profit turning into cash, or merely staying on paper?
- How much capital does growth consume?
- Is management promising more than it delivers?
- What would make this business fail?
Those questions were not obstacles to learning. They were the learning.
Today, AI can compress that work dramatically. Ask it for a company’s competitive advantages, risks, valuation, and outlook, and it will give you a confident-looking answer. The danger is that a polished answer can feel like understanding.
But having an answer is not the same as being able to arrive at it.
Borrowed conviction breaks under pressure
Imagine two investors studying the same company.
The first spends days reading the business, checking cash flows, understanding management’s past promises, and forming a valuation. The second asks AI for a complete analysis and invests based on the summary. Both conclude the stock is undervalued.
A year later, the stock falls 40 per cent.
The first investor can revisit the original reasoning: Which assumption failed? Was the business weaker than expected? Was the valuation too optimistic? Is this a temporary setback or a permanent change?
The second investor has information, perhaps even more information, but may not have the internal map needed to judge whether the original thesis still stands. When prices fall, headlines turn negative, and friends panic, borrowed conviction can evaporate quickly.
Behavioural finance calls this the difference between genuine understanding and overconfidence. Markets do not test your ability to repeat a thesis; they test whether you can hold, revise, or abandon it when circumstances change.
Writing is thinking
The physicist Richard Feynman once objected when a visitor described his notebooks as a record of his thinking. He said they were not a record of thinking—they were the thinking. The work happened on paper.
That insight matters for investors. Writing an investment thesis is not just documentation. It is how you discover whether you actually understand the business.
When you write, you must choose what matters, explain why, and confront gaps in your logic. A vague idea such as “this company will grow” becomes sharper when you must specify: growth from where, at what cost, with what margins, and against which competitors?
AI can help you improve that draft. But if AI writes the first draft and you merely accept it, you may receive polish without developing judgment.
A study on AI-assisted writing found that people who passively copied AI-generated text reported lower confidence in doing the task independently than those who first created their own draft and then used AI to revise it. The practical lesson is clear: use AI after you have thought, not instead of thinking.
What to delegate—and what not to
There is no virtue in doing manually what a machine can do accurately in seconds. Calculating ratios by hand will not make you Warren Buffett. The goal is not nostalgia; it is judgment.
A useful rule is this: delegate information work, but retain judgment work.
| Let AI help with | Keep for yourself |
| Collecting company data | Deciding what matters |
| Summarising long documents | Reading the most important sections yourself |
| Comparing financial figures | Asking why the numbers changed |
| Spotting inconsistencies | Judging management’s credibility |
| Creating a first-pass model | Testing assumptions and valuation |
| Challenging your thesis | Forming the thesis in the first place |
AI is an excellent research assistant. It becomes a problem when it becomes your decision-maker.
A simple discipline for investors
Before buying any stock, try this sequence:
- Read enough of the annual report yourself. Focus on the business description, management discussion, risks, and cash-flow statement.
- Write your own one-page thesis. Why does this business exist? What could go wrong? Why is the price attractive?
- Attempt a basic valuation. Even a rough estimate forces you to connect growth, profitability, and price.
- Then ask AI to challenge you. Ask for bear cases, missing risks, alternative explanations, and flaws in your assumptions.
- Write down what would prove you wrong. This is your future defence against panic and wishful thinking.
This process may take longer. But investing is not a speed contest. It is a decision-making discipline performed with real money, uncertain information, and emotional pressure.
The real risk
The greatest danger of intelligent technology may not be that it gives wrong answers. It may be that we become so accustomed to good answers that we lose the habit of asking good questions.
For investors, good questions are the foundation of good decisions. What do I know? What do I not know? What would change my mind? Can I explain this investment without hiding behind a summary?
Use AI to become faster, better informed, and more thorough. But do not outsource the one thing that ultimately protects your money: your own ability to think.
Because in investing, the effort is not separate from the learning. Very often, the effort is the learning.
