Researchers from the Riken Center for the Brain Sciences have made a fascinating discovery about human perception: they have found that increased confidence in decision-making can be influenced by the characteristics of perceived noise. The findings of their study, which has been published in the journal Nature Communications, suggest that individuals can feel extremely confident in their judgments even when presented with insufficient objective data.
This phenomenon is particularly noticeable when it comes to recognizing familiar objects. People often have no trouble instantly identifying these objects, even if the image quality is poor. For a long time, scientists have been puzzled by the disconnect between the decision-making process and the level of confidence.
Khakvan Lau from Riken commented, “Theoretically, it is assumed that humans are rational, but empirical data often show the opposite.” The study discovered that this confidence is more likely to arise when perceiving fuzzy or “noisy” images. However, the key revelation was the influence of the structure of this noise.
Lau explains, “If the image becomes more visually expressive and noisy at the same time, while maintaining the overall signal-to-noise ratio, we become more confident in what we see, even though our actual visual perception hasn’t improved.”
Interestingly, artificial intelligence that has been trained on similar data has demonstrated a similar bias in confidence to humans. Lau explains, “Once we’ve trained the AI model, we can ‘analyze’ it to gain a better understanding, unlike the human brain.”
This discovery highlights that our judgments and confidence in them can be significantly distorted by the structure of perceived noise, even if the noise appears to be random.