Related Experiment Videos
Humans can consciously generate random number sequences: a possible test for artificial intelligence.
1Department of Neurology, University of Cambridge, United Kingdom. Navindra.Persaud@univ.ox.ac.uk
Medical Hypotheses
|June 1, 2005
Summary
Healthy adults can generate truly random numbers, unlike computer algorithms. This finding suggests potential for artificial intelligence and has implications for understanding human decision-making and free will.
Area of Science:
- Cognitive Science
- Neuroscience
- Artificial Intelligence
Background:
- Computer algorithms generate pseudorandom numbers, contrasting with true randomness from natural phenomena like radioactive decay.
- Human ability to generate random numbers remains an area of scientific inquiry.
Purpose of the Study:
- To investigate the capacity of healthy adults to produce random number sequences.
- To analyze the statistical properties of human-generated random numbers.
Main Methods:
- Participants were instructed to generate and vocalize sequences of random numbers.
- Statistical tests were applied to assess uniformity, independence, and information density of the generated numbers.
Main Results:
- Human-generated random numbers exhibited uniform distribution.
- The numbers were found to be independent and unpredictable.
- Results indicate humans can produce statistically valid random sequences.
Conclusions:
- The findings support the hypothesis that humans can generate random numbers, challenging the exclusive domain of natural phenomena.
- This capability in humans suggests potential for artificial intelligence and neural networks to also generate random numbers.
- Understanding the neural mechanisms of human random number generation may offer insights into decision-making processes and the nature of agency and free will.