That… seems like a silly concern? At best I think you could pull off tricks based on specific timeslots of the day/week (e.g.: lunch rush, weekends, business hours) and most of that was already realistically feasible through bossing around normal humans – no tech required.

Maybe I'm off-base, but it seems to me like the following things would severely hamper the efforts of anyone attempting to do this at scale:

• Exploding the complexity and error surface across all price-setting procedures
• Micromanaging, per location, when/how prices change to compensate for locality-specific stuff like daylight hours and foot traffic
• Avoiding the very legally dubious scenario of prices autochanging in the time gap between taking an item off the shelf and paying at the register

All of that effort in exchange for what? A few extra quarters on the margin if you do it right but losing a few dimes on the margin (or a class action suit) otherwise? I can already see middle-manager heads exploding in contrition. It might happen someday once active AI store management finally oozes into place… but until that day comes I think shoppers can feel safe knowing that they are merely getting gouged in the same old fashion as their forebearers.