Jordan Harbinger avatar

Dexa/Jordan Harbinger

Learn more
Published Jun 8, 2021

518: Daniel Kahneman | When Noise Destroys Our Best of Choices

Nobel Prize-winning psychologist Daniel Kahneman delves into the concept of noise and its detrimental impact on human judgment across fields such as law and medicine, sharing strategies for reducing bias through decision hygiene and the promising role of algorithms, alongside reflections on his Holocaust experiences and scientific career.
Episode Highlights
The Jordan Harbinger Show logo

Topics covered

Popular Clips

Questions from this episode

Episode Highlights

  • Judicial Noise

    Judicial decision-making is fraught with variability, or "noise," which can lead to inconsistent outcomes. explains that judges' decisions can be influenced by factors as trivial as their mood or hunger, leading to significant disparities in sentencing 1. This noise is distinct from bias, which is a consistent deviation in judgment. Kahneman uses the analogy of measuring a line with a fine ruler to illustrate noise: "The measurements in principle should be identical, but they vary. That's noise" 2. Algorithms, while not immune to bias, offer a noise-free alternative, though societal bias against them persists due to a preference for human judgment 3.

       

    Medical Algorithms

    In medical decision-making, the potential of algorithms to outperform human judgment is significant. Kahneman notes that algorithms can reduce errors in decisions like granting bail, where human judges often falter due to noise 4. Despite their advantages, there is a strong bias against algorithms, as people prefer human intuition even when it is less reliable. Kahneman advises caution when dealing with tired doctors, as fatigue introduces noise into their decisions: "When they are very tired, they make poor decisions. And that's noise" 5. This highlights the need for consistent, noise-free algorithmic input in critical areas like medicine.

       

    Algorithmic Bias

    The biases inherent in algorithmic systems and public perception challenges are complex issues. Kahneman acknowledges that algorithms can be biased, often reflecting the biases of their creators, yet they remain noise-free 3. Public resistance to algorithms stems from a preference for natural over artificial solutions, as seen in the discomfort with self-driving cars or AI diagnostics. Kahneman explains, "We define noise as variability that you don't want," contrasting it with beneficial variability in evolution 6. This distinction underscores the need to address both bias and noise in algorithmic decision-making.

Related Episodes