For example: AI writes the contract (or adapts the last used contract for the present need). AI on both sides reviews and makes suggestions. Does AI give the all green signal before humans sign? Not at my company, but I would not be surprised if it's happening.
I see an increasing intolerance for humanless AI communication cycles. It should be identified for what it is, and completely rejected: meaningless.
The silver lining: Humans used to tolerate near meaningless communications of various types written by other humans. Marketing fluff. Blustery corporate exchanges. Distracting technical explanations. When AI writes a thing that has little or no meaning, it feels much more offensive, and more people decide to reject or avoid it. That's good.