It's clear the people adopting the AI systems are the big co's -- we've been battling half-assed phone navigation agents for a decade now, and more recently chatbots (often with humorous results -- see the "No Backsies" car dealership chatbot). AI always favors the one with more resources.
When you have AI's battling AI's, rather than making progress, I suspect it'll actually result in confusion, more delays as tighter roadblocks pop up, and probably some humorous events like the infamous Dominos-vs-Papa-Johns prank call https://www.youtube.com/watch?v=ALnPVybD9X8 but darker because it'll be about cancer treatments instead of pizza prices.
The only winners will be data center operators, and the losers will be everyone without access to the AI tools. In an arms race, everyone loses in the long view.
They already do, and it's pretty obvious.
I've always wondered why don't we have companies that "equalize" that fight, with some kind of subscription or other revenue model that aligns with the consumer. They could also have armies of lawyers fighting on the side of the claimer.
In other words "suspicious bot activity detected, start over" and "account suspended for security reasons" will become new crappy barriers to claims and appeals.