CONCEPT
Uniqueness Bias
The conviction — particularly destructive in Flyvbjerg's taxonomy — that this case is different from all comparable cases and therefore exempt from the base rate. The cognitive distortion that prevents reference class forecasting and that pervades the AI discourse.
Uniqueness bias is the specific cognitive distortion that prevents the outside-view discipline of reference class forecasting from correcting the planning fallacy. The planner insists that this project is different from all previous projects, that the team is better, the technology is more advanced, the circumstances are unique — and therefore the statistical regularities governing every comparable case do not apply. Flyvbjerg has documented uniqueness bias as perhaps the most destructive cognitive distortion in the planning context, because it operates precisely at the moment when comparison with the reference class would correct the forecast. In the AI discourse, uniqueness bias takes the form of the repeated insistence that this wave of AI is categorically different from every previous wave of AI — an insistence that is unfalsifiable and identical in argumentative structure to every previous such insistence that proved wrong.
In The You On AI Field Guide
The bias has deep cognitive roots. Humans perceive