There's a great new "use case" for AI: dodging bait and switch laws! Sure, normally if a dealership employee explicitly offered a car for a given price in writing only to reveal it was incorrect later it would be illegal, but when an "AI" does the same we suddenly can't hold anyone accountable. Ta-da!
I'm not sure exactly how this would play out, but it seems intuitively not true. If I convinced the front service staff at McDonalds to sell me the store for $1, that obviously wouldn't be seen as a valid deal.
The employees do not have the right to sell the store at any price, so I don't think the analogy holds up. From a short bit of googling:
"In Federal Claims courts, the key components for evaluating a claim of improper bait-and-switch by the recipient of a contract are whether: (1) the seller represented in its initial proposal that they would rely on certain specified employees/staff when performing the services; (2) the recipient relied on this representation of information when evaluating the proposal; (3) it was foreseeable and probable that the employees/staff named in the initial proposal would not be available to implement the contract work; and (4) employees/staff other than those listed in the initial proposal instead were or would be performing the services."[0]
Does a support chatbot have the right to reprice items for sale though? Seems like the same situation. Tricking some low level employee or bot in to saying something they shouldn't doesn't seem to be that important.
I also ANAL and I have no clue if it would hold up in court; I was more sardonically drawing an analogy to how section 230 of the DMCA shields companies from responsibility for content surfaced/promoted by "algorithms" while traditional human-backed publications face more liability.
I certainly hope we don't make the same mistake twice!
