In the debate of free will vs. determinism, we can make the connection between computer neuro-network architecture and organic brain biology. We observe the AI’s impressive and capable performance, producing large amount of useful output when there is a large volume of quality training data. The output from this “well trainied” AI is valuable, and in turn the AI is valued.

Conversely, we observe impressive productivity and the production of useless output when the AI receives a large amount of poor quality training data. This AI’s output will be generally regarded as having little to no value, as then does the AI.

The exact thing occurs in the human brain. When people are exposed to poor training data, they produce useless output, and lessen their utility to near zero, and realize reduced empathy from others.

So where free will perhaps can most fully be expressed is in the careful selection of the training data to which we choose to expose ourselves.