Random numbers are used all the time in computer science. My background is mostly scientific and engineering where random numbers are used quite often for Monte Carlo simulation techniques. The domain data space is too vast to exhaustively test all the possibilities so in order to explore solutions the simulation randomly distributes the values of the input parameters generating a large number of output results which can be validated with various statistical tests. For instance you can calculate the value of pi using random numbers in this way.
So, it is critical that random numbers not only be jumbled and without any apparent order, but they must be equally likely to occur (flat distribution) in order for any Monte Carlo technique to work properly.
Most random number generators provided as a part of whatever computer language you're using is not really random. They are software algorithms and as such they are deterministic. These random functions we use are usually referred to as pseudorandom number generators. And they have been engineered to do a reasonably good job producing a flat distribution of jumbled numbers.
But is that enough? I will address that in Part 2
Is this bias acceptable? Will the answer get better with another 12 hours? These are random questions for you to consider.