March 16, 2012      

Apparently, a fair amount of talk at this week’s European service robot gathering, Innorobo, concerned whether or not a Jetson’s Rosie or a Forbidden Planet-like Robby companion robot is a dream worth pursuing or even possible in the foreseeable future, prompting the above headline to appear atop a BBC article about the event.

True, the technological challenges might seem daunting, and the field of AI has historically been rife with disappointments and false starts. But two unrelated technologies have emerged recently, which could make multitask-capable service robots with appealing personalities happen faster than many suspect. They include:

  • The rise of ultra-fast, semantically aware computers such as IBM’s Watson, which could provide cloud-based knowledge support to fleets of robots, remotely.
  • The even faster rise of intelligent agents modeled after Apple’s Siri. Think of such agents not so much as apps able to give instructions, but as apps that conceivably could mine the experiences of millions of users and the robots they own.

Ergo, if one robot in Peoria and its owner figure out a more efficient way to fold the laundry, that knowledge would become accessible to every other robot connected to the same cloud — instantly.

What’s lacking at the moment in order for all this to happen is a large enough installed base of robots to provide the data input — that is, the relevant user/robot experiences so robots can effectively teach one another how to do things.

But that should change as Siri-like apps become more versatile in the assistance they provide. Perhaps in a few years you’ll be able to ask the best way to cook a turkey. Your agent/robot will know the type oven you own. And it would know the size turkey you bought and its likely condition based on when you removed it from the refrigerator.

All a robot needs know beyond that in order to prepare a turkey is how to carry it on a platter and place it inside the oven. And some family in Peoria has probably already taught their robot to do just that, making it possible to pass that knowledge on.

Yes, the above brings up a bees’ nest worth of privacy and perhaps even liability issues. But the point is the technology will get us there sooner rather than later. The dream, in other words, remains very much alive.