While I would certainly prefer to be able to pit my expert systems against theirs, rather than beat my own head against the wall; but that sort of situation strikes me as a deeply perverse arms race: if you have an expert system capable of wasting vast amounts of my time in the course of providing ‘service’; and I have one capable of fighting the battle of attrition necessary to receive that ‘service’, it sounds like we have a problem that our respective bots could hammer out in a matter of seconds(depending on latency) if they were actually seeking a solution.
This doesn’t mean that I would hesitate to deploy a bot in such a situation, if one is available; but wherever possible I’d prefer that their bot just skip the bullshit and be helpful(a few years back, this particular debate seems to have played out behind the scenes in airline ticketing/check-in kiosk systems: some people were drinking the kook-aid on “agents”, hideous hybrids of Clippy and the low-rent talk show hosts of the uncanny valley, that were supposed to make interacting with a computer feel more ‘personal’ or some nonsense. The sane, decent, faction went with systems that skipped pretending to be helpful; and just allowed you to verify times/gates/etc. make a seat selection, and be on your way.)
