· Blog
What gets lost when AI summarises the world for a leader?
Strategy lives in rare information. What happens when the tools leaders use are built to favour the common?
Ask a large language model about your market and it will give you a fluent, balanced, well-structured answer. That is precisely what makes it useful. It is also what makes me uneasy.
These systems are trained to reproduce the most probable continuation of what they have seen. In most tasks, that is a virtue. In strategy, it may be a trap, because strategy is often made of what is not yet probable: the weak signal, the contrarian analyst, the customer behaviour that does not fit the pattern.
The questions I am working on
- Where in the strategic process, searching, framing or weighing options, does AI exert its strongest influence?
- Which of these risks are old problems in new clothes, which are amplified, and which are genuinely new?
- If the risks build on each other, what does that mean for when, and how, leaders should intervene?
I approach these questions by bringing together a large and fragmented literature and reading it through the lens of strategic cognition. My interest is less in listing risks than in understanding how they travel: how a small distortion at the start of a decision can become a large one by the time it reaches the boardroom.