In 2005, "Paro," an interactive robotic pet, was introduced in Japan. It looks like a baby seal, and has 12 tactile sensors in its fur, touch-sensitive whiskers and a system of motors that silently move its parts. While this "pet" didn’t really catch on in Japan, it’s made its way to the US and ended up in some nursing homes.
Paro has been described as a low-maintenance alternative to dogs and cats for pet therapy. It’s an intriguing idea. There are certainly some appealing aspects, in that a robotic pet won’t bite or scratch, doesn’t poop, doesn’t need to be fed, can’t be injured, doesn’t have a large population of resident bacteria in and on its body and can’t become infected with various microorganisms from patients. Those are appealing from an infection control standpoint. However, despite this, you can’t approach Paro as a way to eliminate infectious disease risks, since the robo-pet could easily become contaminated by someone, then spread infection from person-to-person. For example, if someone has a bacterium on their hands and they touch the robot, they could transfer the bacterium to its "coat." It could then spread the bacterium to the next person that touches it. Infectious disease risks would be lower, but not zero.
Pet therapy is all about cost-benefit. We know there will never be a zero-risk pet-human interaction. However, socializing with an animal can provide significant benefits to many people, and I’m not convinced that the same degree of benefit would be provided from interaction with a robotic critter. There might be some situations where a robotic pet would be useful in a nursing home or similar environment, but I don’t think they’re going to replace interaction with a live animal.