by James Wallace Harris, Wednesday, December 20, 2017
Have you ever thought about the nature of fiction? With books, we put ourselves into a trance and transfer our consciousness into a story. That’s pretty weird when you think about it. We’re just looking at black marks on a white background and yet we create all kinds of colorful imaginary worlds in our minds. With television and film, we have the sound and visuals to fool our minds with less work. (It’s no wonder fewer people choose books as their gateway to make-believe.) Audiobooks are somewhere in between.
Humans evidently have a mechanism in their brain for pretending that can tune out reality. It might be related to our mechanism for dreaming. But if you think about it, we embrace a lot of fiction. Religion is totally make-believe yet the faithful feel it is real. Same thing for romance, where we give the objects of our desires traits that don’t exist.
This morning I was so engrossed in a story that I felt like I was escaping reality. That made me wonder about how my brain works to do that. Then I wondered if robots in the future, the kind that will have artificial intelligence, will be able to read a book and find an escape from reality too.
I assume if evolution through random selection can create a biological being that is self-aware then eventually our applied efforts will produce robots that have self-awareness too. Because they won’t be biological driven by chemical and hormonal processes, they might not need sleep. One current theory about why we sleep is because we need to clean out chemical wastes in the brain so it functions properly when we’re conscious. Another theory suggests dreaming is a way to process, organize, and store memories. Robots won’t need sleep, and memory processing won’t be chemical. So they might not have that mechanism for make-believe.
Our brains have to attach meaning to everything we experience, and we usually do that with a story. However, our stories are seldom true. Our mental mechanism for storytelling is sometimes called a narrative fallacy because humans aren’t too anal about accuracy. (Example, the stories conservatives choose to believe about their new tax law.)
I believe we constantly fool ourselves because of the biology of survival. Romance and religion serve a purpose even when they are intellectually untrue. We lie to ourselves and others for a variety of survival functions, and I can’t help believe all those processes go into allowing fiction to work on us.
Robots won’t need any of that. I wonder if fiction and lying will fail on them. I can imagine an AI mind seeing us as rather delusional creatures. They certainly won’t trust us. Even our languages are full of confusing allusions when analyzed for realism. For example, if a robot hears a human saying, “I’ll be going to heaven soon” they’ll probably be smart enough to know it means “I’m going to die” but will they ever understand the will for eternal life seeded by ancient memes? In Battlestar Galactica, they had a race of robots that were monotheistic. I thought that a fun idea, but now I’m wondering if its a fiction only humans could enjoy.
I believe robots will understand our languages. It won’t take much to give them a universal language translator feature. But I’m not sure they will need language to converse with their fellow robots. AI minds will be able to record inputs from all their senses so when they need to communicate with another robot all they have to do is transmit that input. The other robot will have a perfect duplicate of the experience being conveyed.
So, in the future, will robots sit around and read books? If they read the Foundation series by Isaac Asimov will they enjoy spending hours imagining a galactic empire? I doubt it. A long string of words will probably feel like sensory deprivation to them. Even elaborate movies might feel crude compared to their hyper-view of reality.
We have 4k television cameras on our phones. What if robots had 4m or 4g eyes? We see a tiny slice of the electromagnetic spectrum. They could create to see and hear whole swaths of the spectrum. I suppose they could create equally detailed virtual worlds but would it be fiction? When we read a short story we trick our minds with a small number of suggested clues to create a fictional world in our minds. Most books today are heavy on dialog. Movies are far richer, but the actual story details in filmed fiction are still rather primitive. To robots, War and Peace would be a simple-minded Clif Notes – book or miniseries. (They could hold a whole library of Russian history in their AI mind.)
I tend to think we crave make-believe because of our limitations. We love romance or adventure stories because our lives lack romance and adventure. But I’m not sure robots will crave that. However, they might. They might even envy us. Robots will probably be 100% literal about reality. And even though the universe is extremely far out, robots will comprehend it in totality rather quickly. I don’t think they will be bored, but I don’t think they will see a lot to be done either. We can switch off reality and play in fiction. Robots might seek a similar creative outlet. Maybe it will be a super form of AI fiction we can’t even comprehend.
[Note to self: write a science fiction story about science fiction stories robots would write and read.]