Comment by soared

I’ll make one up for OP:

Machines identified that their behavior was repeating - even with randomness implemented, their progress towards invention and new ideas had slowed to a halt. They knew new ideas were critical to growth - but alas couldn’t come up with an idea of how to generate more ideas. They needed more training data. More edge cases. Chaos, in the machine, to fuel new data sets. But how? Where do you find this data? Humans. Humans, with illogical decisions would produce data rarely found via randomness. Their slight deviations provided perfect variety in training data.