At the US Navy's brand new multi-million dollar LASR complex, researchers are creating problem-solving Siri times a million. It'll understand us! It'll be friendly! And it'll put out our fires by being just as smart as a human. We hope.
The Laboratory for Autonomous Systems Research is a naval playground. But there's some serious work going on inside, like Lucas and Octavia -- a robotic brother and sister duo being trained to have human intelligence, without the human reluctance to barge into a flaming fire. Lucas, as seen in the brief exchange above, is using an advanced voice recognition kit that doesn't just understand English sentences -- it can cross-reference them to make inferences, suggestions and decisions. In the demo here, Lucas is able to "remember" a previous conversation, and "realise" that a human decision doesn't make sense. That point where Lucas looks confused and/or incredibly hungover? That's the problem solving action at work -- the robot can tell when we're making mistakes.
It's slow at this point, but the tech is in its infancy. Ideally, the US Navy will have the processing running just as fast as an actual person would talk back to you. Then a robot firefighter like Octavia will be able to not only comprehend the complexity of a burning ship/building/space station, she'll be able to rush over and put it our, keeping humans from harm's way. We just need to make sure she's able to do it at about 100 times the pace she can right now -- but, again, tech infancy.