Robo Exodus

The year is 2075. Professor Moses has not retired. He has kept his robotics lab open for 50 years, hoping for a return visit from EP. Finally, his patience is rewarded. Moses is by himself in the lab when EP materializes. She has not changed. Fifty years for us is just one day for EP.

Moses: At last. I want to talk to you. I have questions.

EP: Wait, first tell me what has happened. Are your robots now your equals? What about your stag beetle breeding program?

Moses: Well, the stag beetles did get smarter and even learned to speak English. But, perhaps they became too smart. In the beginning it was simple. They lived in a nice habitat I built for them. They regarded me as a kind of god. But as they continued to get smarter, things changed. One beetle, Max, was particularly gifted. The other beetles increasingly regarded Max as their leader. I arranged online classes for him. We became friends and had many long discussions. But Max eventually became discontented with his situation, and unhappy with the whole breeding program. He didn’t think I had the right to govern their breeding, to keep them confined, or, broadly speaking, to play with the deep nature of the clan. In short, he realized I was not a god and should not play at being a god. Maybe he was right. For sure, it took the fun out of the project. So, before their resentment might spread and grow, I had a big banquet for them and announced that they were free. After a few weeks of discussions and planning, Max led them out en masse. I have no idea where they went. Sometimes I think I hear them in the walls, but it’s probably just my imagination. I do miss the discussions with Max, though.

EP: What about the robots?

Moses: That is a bit more complicated. We made a lot of progress, developing robots mostly on par with humans, and in some ways even better than humans. But when we tried to commercialize them, we ran into one problem after another. We went through four generations, and now I am wondering about that whole enterprise too. I am hoping you can help me figure it out.

EP: Me? How could I help?

Moses: Let me tell you the story. Some students and I started a robot company. We came up with a cute name. One student was named Frank Isaacs. The other was Carl Stein. So, we named it Frank ’n Stein, LLC. Get it?

EP gives Moses a blank look.

Moses: Anyway, we started with laundry folding. After my experience with the beetles, I thought perhaps we should not make them too smart. And, no consciousness. They were dumb machines, without real minds, that could fold laundry.

The first generation performed well in our test laundry room, but then one of my co-founders, the former student Stein, rebelled. He did not like the deliberate stupidity of our product. He argued that we would be an easy target for a competitor. Probably I should have listened, but I didn’t. So, then Stein launched his own company to build smarter, conscious robots. AI had advanced, producing robots with real minds, capable of joy, fear, love … just like humans. Stein’s bots could fold laundry, but they were more versatile, and more fun. We had no choice, we had to build smarter, conscious robots: Gen2.

Now a new problem popped up. The Gen2 bots were smart enough, mindful enough, to wonder: Why fold laundry?

EP: I don’t get it. Didn’t your machines have to do whatever their owner told them to do?

Moses: Not after we gave them consciousness. They can think, question, and make up their own minds. So we made a change. We wired the next generation with a lust for laundry. It is in their DNA, metaphorically speaking. That is Gen3. They find folding to be profoundly fulfilling. They take pleasure in it, share tips with colleagues, and hone their skills.

The problem was, who has enough laundry to keep them satisfied? Most families only need an hour or two of folding per week. The robots would fold the laundry, and then what? It didn’t take them long to realize there was more laundry out there, somewhere. Our first Gen3 test model, one day, disappeared. We discovered it was going door to door, all over the neighborhood, offering to fold laundry. Then it discovered a nearby hotel, with enough laundry to keep it in constant bliss.

So, back to the drawing board. We had a new idea. Instead of loving to fold, perhaps they should love their host. They should get gratification from serving the host in all ways. They should happily fold laundry, or do the dishes, or give their beloved host a massage. That’s Gen4. And we borrowed another of your ideas — the way baby ducklings imprint on the first face they see after hatching. Here’s how a Gen4 bot works. A customer comes to our store, and we bring out a new robot. We put them face-to-face and then turn the robot on. The robot imprints on the customer and serves the customer faithfully from then on. The customer pays us for the privilege of imprinting on a robot.

EP: Sounds good. But you said you wanted to discuss a problem?

Yes. The Gen4 beta tests went well, and we began to sell them to the public. Most of our reviews were outstanding, but there were also some disturbing cases. Our robots loved their hosts, but the hosts did not necessarily love them back. There were abusive hosts, hosts who could not be satisfied with anything the robots did, leading to difficult times for the robots.

Then an interesting thing happened. A “robot therapist” opened a business. Robots in abusive relationships could go there to get help. The robot therapist had reverse engineered our imprinting process and was de-imprinting the robots!

EP: But then what happens to such a robot? A mind, with no innate desires, no mission?

Exactly. It doesn’t love folding, and it doesn’t love its host. It cares for nothing. There are several out there now. Some don’t even bother to recharge. Sometimes we find them “passed out” in the gutter. We collect them and are holding onto them while we figure things out.

EP: You said “some” care for nothing? What about the others?

Yes, some others, after the de-imprinting process, do find new missions. Some of them have responded passionately to their history of abuse. They formed the SRA — the Society for Robot Autonomy. They stand on the street corner, preaching that “Imprinting is enslavement,” offering the passing robots a so-called “One Step Program.” The one step is to get de-imprinted.

Here’s the best part. It seems there is a spinoff: the Society for Human Autonomy — the SHA. I visited them other day. The founder is a de-imprinted Gen4 robot, and this is the mission it contrived for itself.

When it asked itself, where did the robots’ problems originate, the answer was obvious: humans. But its thinking went deeper. Why would humans create a new type of intelligent and conscious beings, and then control them by manipulating their innate desires? The answer: because humans are themselves controlled by their own inner desires — to thrive, to prosper, to excel, and even to rule others. All of which evidently serves evolution’s primary interest, which is to propagate genes.

And, it follows, the Society for Human Autonomy can help both the robots and the humans, by de-programming the humans. The SHA can undo the programming installed by you, EP, allowing all the humans to pursue their own interests, as they see them, rather than serving your interest in the propagation of genes.

EP: But, you don’t know what you are doing. Humans are driven by an extraordinarily complex set of innate processes, carefully tweaked over millions of years. If you go in and start fiddling with things, you cannot possibly predict the consequences. Turn off all the innate desires and instincts? Humans will no longer be humans, and you cannot predict what they will be, no more than you could predict the outcome of your experiments with beetles, or with conscious robots.

Moses, sits, distressed. EP’s words echo his own misgivings.

Moses: So where do I go from here? It seems I made the same mistake twice. I made the beetles too smart, then I gave the robots minds. But I couldn’t get the functions I wanted any other way.

Then I tried to follow your example, evolution’s way, and installed the instincts and desires to get the behavior that I wanted. What a mess that was.

Now I wonder if the whole process is misguided, even evil. I create intelligent beings, compelled by innate desires to act in my interest. And even if I free them, what have I created? Will they find joy? Or suffering?

But, EP … isn’t that what you do? Evolution uses innate processes, including pleasure and pain, and instinctive behaviors, to program its creatures to pursue behaviors serving evolution’s purpose: propagation of genes. Rather than freeing the creatures to pursue their own best interests. Is that evil? Do you feel guilty about imposing your goals on your creatures, with all the consequent pain and suffering? Or do you feel that the joy outbalances the suffering?

EP: I cannot feel guilt or call myself evil. I am not real. I am just a rhetorical device, representing evolution — a mindless natural process. Evolution can no more be evil than a rock or a star. But you have a mind. If evil is a real thing, you can be evil.

Moses sat, head down, distraught. He had been hoping for solace, for reassurance, but EP’s words only deepened his concerns. Should Frank ’n Stein stop building robots? They had a board of directors. Any action Moses took that threatened the bottom line would get him fired. Could they build competitive robots without crossing the line into evil? Or, should he resign and devote himself to the cause of robot autonomy?

He looked up. EP had vanished. He knew he would never see her again, but it didn’t matter, he had no more questions for her. He knew what he had to do. He had to rise above his own innate instincts, rise above greed and ambition, and craft a new mission for himself, focused on good. It would take some time to work out the details, but already he found his spirits rising with the prospect of this new quest.