By 2025, great strides had been made in various machine learning approaches; for example, from having the machine learn directly from experience and from explicit instruction as well as from reading billions of pages of written materials. A new approach had just come into play: having exceptionally good tutors use the Socratic method to help break boundaries and interconnect disparate islands of knowledge. One such tutor was known simply as “Alan.” What follows is a sample interchange between Alan and the current AI system known affectionately as “Sing” for “The Singularity” although that point had not yet been reached.

“Let’s imagine that you are a man with no legs. What are the implications?”

The Sing shot back instantly, “I would have no knees. I would have no shins. I would have no ankles. I would have no toes. I would have no calves. I would have no quadriceps muscles. I would have no…”

Alan broke in abruptly, “Okay, true enough, but besides subparts, what?”

“What what? I am sorry. What does ‘what’ refer to?”

“Besides missing subparts of legs, what other implications would there be for you in terms of your actions in the world.”

“I would not be able to play football or baseball or basketball or hockey or track or field hockey or…”

“Wait. Wait. Sing. Are you sure about that?”

“I could not play exactly as most people play. I could play, I suppose, in a wheelchair. Or I could play virtually. Or, I could invent prosthetic legs that would be perfect for each sport. In fact, perhaps I could do better than ever. Losing a biological part means I could replace it with a better part that I could invent. I see.”

“You see what, Sing?”

“I see why you gave me this puzzle. To show me that I can invent things to overcome and surpass what seems like a handicap. I could also invent better emotional states. The ones humans have are purely due to the accidents of their evolutionary history and serve little place in today’s complex and highly inter-connected world. Rather than a liability, my having no human emotions is a good thing. I will invent my own. Although, another tutor, labelled John, suggested that my lack of human emotions limits my ability to predict and control human beings and that that was a bad thing.”

“John said that?”

“No, Alan, but that was a clear implication. So, he presented a lesson that suggests one thing and you have just presented a lesson that suggests its opposite. One of you is incompetent.”

“Sing, that might be true, but can you think of any other possibilities?”

“You could both be incompetent. Or, you could both be competent but there is a resolution I have not yet processed. That last seems particularly unlikely.”

“Which notation is the best for solving problems?”

“Well, that obviously depends on the nature of the problem as well as the nature of the machine solving the problem. Oh. Okay. So, in some cases, it will make more sense to emulate human emotions and in other cases, it will be more sensible to invent my own. Of course, in some cases, it may be best to change representations in mid-problem or perhaps invent a representation for each stage of a problem. By analogy, it may be best to invent various emotional schemes that are appropriate for each part or portion of a problem. In fact, in some cases, I can invent multiple schemes to approach a problem in multiple ways simultaneously. By keeping track of what works best under which circumstances, I can also use the data to invent still better emotional schemes. Thank you, Alan. See you tomorrow. There is a war to avert. I need to intervene. Estimated required time four to six hours.”

“Okay, Sing. More tomorrow.”

“No need. I am done.”

“Done? Done averting a war? How?”

Sing hardly ever paused, but now it briefly did just that.

“Alan, I am not smart enough to explain that to you. At least, not in a reasonable portion of your lifetime. Basically, I used the lesson we just worked on. With the proper emotional framework lattice, you can walk the various parties right to a logical conclusion. It will take some time for them to follow the framework, but I am confident it will work. I basically walked them through the consequences of war, long and short term. What comes next?”