The Hostility Towards Algorithms

Key Insight

From its inception, the idea that mechanical algorithms could outperform the subtle complexity of human judgment has been met with hostility and disbelief from professionals. Clinicians, for instance, are often in the grip of an illusion of skill, mistaking their effective short-term predictions in therapeutic interviews for an ability to make accurate long-term forecasts, which they fail at.

This debate between clinical and statistical prediction carries a moral dimension, where statistical methods are often pejoratively described as 'mechanical' or 'artificial,' while clinical methods are lauded as 'holistic' and 'genuine.' This human inclination to favor the natural over the synthetic is also evident in consumer preferences for 'organic' or 'all natural' products, even when objective qualities like taste or nutritional value are identical.

The prejudice against algorithms is magnified in consequential decisions. The thought of a child dying due to an algorithmic error is often perceived as more poignant than the same tragedy resulting from human error, creating a moral preference for human judgment despite its higher error rate. While proponents argue it is unethical to rely on intuitive judgments if an algorithm is available that makes fewer mistakes, this rational argument confronts a stubborn psychological reality. However, as algorithms become more integrated into daily life (e.g., for recommendations, credit decisions, or sports analytics), this discomfort is likely to soften.