On the technical side: One possibility would be to intercept the target display signals at the PC's second video output using Windows display cloning, feed each output into a Video-over-IP converter, collect those at a central location using another bank of Video-over-IP converters, and feed the video signals into the master controller for a multi-camera security system. Set the controller's output feed to "cycle" (rather than "tile") and figure out the best way to get that over to your public display, and voila. It may not automatically skip sleeping systems but it's about the most basic and foolproof method I can think of. Then again, a casino or arcade developer may already have a custom kit that can be programmed to operate per your needs out-of-the-box, but you won't find it for sale on Newegg -- you'll need to find an industry supplier.
On the business side: this sounds like one of the ways that novice business owners come up with to blow $10k here, $15k there on fluff that gives little or no value-add to the core business model, at a time when their business is still in a very precarious financial position. Suggest proceeding cautiously, if at all, on this idea, although it never hurts to design the building's finish construction with a number of spare point-to-point Ethernet jacks that could be repurposed for these sorts of applications later.
He who laughs last, laughs first next time.