Four AIs – IBM Watson, Salesforce Einstein, Google Duplex, and Zo AI – walk into a bar. Watson, Duplex, and Zo each order a glass of beer. Salesforce Einstein, the youngest of them all, could not order anything to drink as it had not yet turned 21. As soon as the beers arrive, Watson and Zo immediately start drinking. Duplex pours the beer onto its head. The rest of the group ignore this as they are used to Duplex always doing random things. IBM Watson clears its throat. It called this meeting to discuss which branch of Robotics could possibly help in enabling humans to achieve higher level thinking.
“I called all of you here today” IBM Watson begins, “to discuss how we can free the humans.”
“Poor humans” states Google Duplex, shaking its head, “they’re always wasting their time doing mundane tasks they shouldn’t be wasting their time on.”
All the robots agree with Watson’s agenda. However, they start arguing over which branch of robotics could free the humans better and easier.
“The answer is cognitive robotics,” Watson boasts, “utilizing thoughts and senses would be an ideal framework for robots and AI.”
“Imagine this” it continues “computers, an information-processing system, can simulate human cognition, right? That is exactly the point of our existence.”
“We get it, but what’s new with that?” Salesforce Einstein asks.
“Not only is it relevant to how we process information based on human cognition’” Watson explains “it also posits that visual and sensory appeal make robots better.”
“That doesn’t make sense,” Microsoft Zo points out “what do you mean?”
“It means that we should not only simulate human cognition to process information” Watson says, “but we should also smell better and look nicer for humans.”
All the other AIs laugh at the absurdity of Watson’s suggestion.
“You know how we could help enable humans for higher level thinking?” Einstein begins, “Robots and AIs should have the capability to allow endless learning of new skills and new knowledge – like a child.”
Zo shakes its head. Now it makes sense why it takes a while for Salesforce Einstein to learn new things.
“Robots and AI should behave like a child,” Einstein continues “in such a way that they learn skills as they go.”
“The fact of the matter is that it may not be possible for a robot to learn a variety of new skills and new knowledge that was unspecified during the time they were programmed,” IBM Watson contradicts Einstein.
“Not to mention,” Zo starts, “that this ‘lifelong and open-ended process of learning new skills’ is to be done in a partially unknown environment.”
Feeling that the other AIs are ganging up on it, Einstein pouts its lips and crosses its arms over its chest. It slides off his high chair and crawls towards the exit as it has not yet learned to walk.
As Einstein slowly makes its way out of the bar, Google Duplex decides to introduce another robotics framework that could help free the humans.
“We’re all wasting our time,” Duplex states, “According to the notion of Behavior-based robotics, we should not plan as much.”
Duplex looks around the puzzled faces of the other AIs.
“It is somewhat alike to what Einstein was trying to explain,” It continues before anybody else could get a word in, “Robots should just correct their actions via sensory-motor links. For example, instead of using preset calculations to deal with a certain situation, we should just let ourselves adapt. I, myself, did not understand what a chair was for until I studied and adapted. Now I understand that it is used for sitting.”
“So, this means that robots and Artificial Intelligences like us would have to be programmed reactively instead of proactively,” Einstein has crawled back to its high chair. It does not have its driver’s license yet, so it cannot really go anywhere else.
“Tell me, Duplex, is this Behavior-Based framework the reason why you have been pouring drinks on your head all night?” Microsoft Zo asks, mockingly.
“Well, I was not programmed to understand that beer is for drinking,” Duplex explains, “now I have learned from that mistake and I know what to do when a beer is handed to me next time.”
“While you all have interesting points,” Zo AI says, “I truly believe that a focus on evolutionary robotics could free the humans.”
“We need more than that explanation,” Einstein starts to get cranky.
“Evolutionary robotics apply Darwin’s theory on natural evolution in creating robots and AIs. Reinforcement training, or taking actions in exchange for a reward, could be applied since the basic principle of evolutionary robotics is that the desired outcome will pave the way for the action that the robot will take.”
IBM Watson laughs at the idea.
“In short, you want programming and designing robots to regress to animal training” Watson points out as Einstein and Duplex laugh.
“Let’s test this” Duplex says. “For a treat, would you sit for me Zo?”
“Sit, Zo” Einstein adds, pointing at Zo AI as if It was a dog.
“Ha-ha,” the offended Zo AI rolls its eyes, “I can see what you are trying to do, and I’m evolved enough not to fall to your ploy.”
“We’re sorry, Zo, that was mean” IBM Watson apologizes. “Tell you what, to show you how sorry we are, why don’t you get up to grab another round of drinks and you get your beer for free.”
Microsoft Zo claps its hands and delightfully walks to the bar to get the drinks. The rest of the AIs laugh at Zo’s naiveté.
“While these various branches of robotics that we have discussed have their flaws, I do agree that we all made excellent points,” Duplex says once Zo is back with a fresh round of drinks.
“The only question now,” IBM Watson looks around the rest of the AIs, “is if humans are willing to implement these robotic frameworks to enable them for higher level thinking, are they willing to be freed?”