[jschveibinz]

Serious question: can we define the top 5 parameters that distinguish human-generated non-fiction writing from AI "slop" writing (I don't like the term "slop" because it carries a built-in bias)?

Since AI output is based on the most likely next word/token given the context, let's consider how humans do this, i.e. choose the next word based on context.

Is one of the parameters the degree of "randomization"?

Are these other parameters?

- Specific experience?

- Recent exposure to relevant materials, i.e. paraphrasing?

- Borrowed knowledge from a specific resource?

- Writing style, i.e. choice of words or sentence structure?

I'd like to hear what others think.