
A method based on quantum physics produces strings of numbers that are truly unpredictable.Credit: Flavio Coelho/Getty
The outcome of quantum experiments is intrinsically unpredictable. Now physicists have combined that feature with blockchain techniques to generate random numbers in a fully transparent process for the first time1.
Public sources of random numbers are used for various applications, such as lotteries, jury-duty selection and the assignment of placebos in clinical trials. A process that not only produces numbers that are truly random, but that also does so in a trackable, verifiable way can add an extra layer of trustworthiness, say the researchers who developed the new system.
This is the fastest random-number generator ever built
Their approach builds on a quantum-physics-based technique for generating random numbers that was first demonstrated in 20182 by physicists at the National Institute of Standards and Technology (NIST) in Boulder, Colorado. It uses a device that produces pairs of photons that are entangled with each other, meaning that they share a common quantum state. The photons in each pair are sent to two measuring stations around 100 metres apart, where their polarizations are detected to produce a string of digital bits (0s and 1s).
The bits are truly random because quantum physics dictates that the photons are not in a well-defined polarization until they are detected, and that the outcome of that measurement is random. Crucially, the way in which the detections are done is chosen at random, and independently, in each location: any coordination between the two locations would require signals to travel between them faster than the speed of light. The results are then independently checked by devices at the University of Colorado across town.
Tamper-proof process
In the latest upgrade to this system, described in Nature on 11 June, the team has made it possible for third parties to verify the process by time-logging every step in the measurement using blockchain techniques. The change means that anyone who accesses the publicly available data would be able to see whether the process had been tampered with.