Kleshas as AI Bias and Distortion

Patanjali identifies five kleshas (afflictions) that distort perception: avidya (ignorance), asmita (egoism), raga (attachment), dvesha (aversion), and abhinivesha (fear of death). These psychological patterns manifest directly in AI systems as systematic biases and distortions. Avidya appears as training data gaps that create blind spots. Asmita manifests as overconfidence in model accuracy despite known limitations. Raga emerges when systems are trained to optimize for metrics their creators are attached to, distorting representation. Dvesha causes aversion to unexpected data or contradictory information. Abhinivesha appears as systems clinging rigidly to established patterns, fearing the death of their learned categories. By naming these kleshas within AI design, we transform them from invisible biases into recognizable patterns we can systematically address. This framework enables developers and knowledge workers to audit systems not just for technical accuracy but for the psychological distortions that undermine wisdom.