I was doing some grocery shopping in Sainsbury’s with my 8-year-old Daughter, Lu, a while ago. I was just heading down the pharmacy aisle on my way to Milk, Cream and Yoghurt. We needed full fat Greek yoghurt because I was doing my famous lamb biryani for dinner that night. We must have had someone coming over because we only do it on special occasions. We’ve done it for Christmas dinner a few times just to give Christmas time a bit of a kick. What can I say, people seem to like it.
Anyhoo I digress.
Yes, I was on my way to the yoghurt when I heard Lu speaking quite rudely to someone. Now we have made a point of raising her to be a very polite little girl, and whilst admittedly Lu can be something of a blunt instrument at times, she’s very chatty and generally speaking she is very polite unless she is fed up with who she is talking to (“Why don’t you just shadap Fatty Daddy”). She’s a one that one.
Anyway, whoever she was talking to was really getting it.
“You stupid thing, why can’t you answer a simple question”
I was taken aback.
“Oi oi, who are you talking to madam?” I admonished.
“Hey Siri, tell me a joke” she commanded.
The immediate sense of “God that’s so rude, what happened to please and thank you?”, didn’t diminish, as I realised, she was talking to my phone. In fact, I was every bit as disappointed in her as if she had been talking to a human being. I was surprised. So, I stopped right there, next to the indigestion tablets, and turned to Lu.
“Erm, Lu?”
“What?”
“I dunno why, but I just feel that you are being really rude to Siri. Do you think its ok for us to be rude to her just because she’s an AI?”.
“Huh?”.
The thought that came to me originated elsewhere, and I said to Lu:
“Whether we are based on carbon or on silicon makes no fundamental difference; we should each be treated with appropriate respect.”
Arther C Clarke 2010: Odyssey Two
What followed was a discussion, about the nature of our future relationship with Artificial Intelligence entities.
Lu is going to sort out Fusion, right after she has met nerdy Brian, who keeps pushing his glasses up on his nose (in like a really cute way?), during her Quantum Physics Masters at Cambridge, so I felt having this conversation was just a relaxing meditation for her, a musing if you will. We covered a lot of ground. Not physically. We only managed to get from Indigestion tablets to Fishermen’s Friends during our discourse, but philosophically, we climbed Everest. Actually, that doesn’t really cover that much ground does it? We crossed the Sahara!
If you were walking through Hyde Park, would you kick a tree? Just walk up to it and kick the living shit out of it? Why don’t we see people doing that? How come we don’t just all spend our time walking from tree to tree screaming abuse and kicking the bark off those tall, leafy bastards? What would that say about us? I’m pretty sure it would say we were bad people.
Maybe we don’t do it, not because we don’t want to, but because we wouldn’t want anyone in the park to see us doing it. In this age of environmental “awakeness”, it would be unacceptable. People would run over and tree-kick-shame us shouting “how would you like it if I just walked up to you and kicked you?”.
“Leave that poor innocent tree alone you horrid person!” they might exclaim. They might even swear and threaten you with physical violence. Look what a hard time we are giving the people burning down the Amazon! Such is the level of empathy we are supposed to have with the environment these days.
Acid question though. Would you kick, or shout at, the tree if you were all alone in a forest and no one could see or hear you do it? I don’t know that I would, you know. I think I’d be ashamed. I’d feel guilty. I think most people would. Which is weird when you think about it. Our current understanding of trees suggests that the kick or the insult itself wouldn’t be “felt” by the tree. It’s not going to spend the next few days wondering what it did to deserve that kicking from that hiker. It’s not sensient, but it is alive, and generally speaking we give it due respect.
Is being alive the quintessential attribute of “being” though? Loads of stuff on Earth is alive but I wouldn’t say they are all beings. I don’t think of amoeba’s, or phytoplankton, or daffodils, or flu virus or trees as Beings. So I don’t think it is just about being alive. Descartes said “I think therefore I am”? So is it thinking that makes us beings then?
Meet Bob. Bob, is our family dog. He’s so cute. Yes you are! Yes….you….are! Look at dat face. We love him to bits because he is so cute, but also because he is the cleverest dog ever! Not because he does tricks (he can do a few), but because of how much he understands and how much he communicates. He solves problems. He won’t come in after he has done his last pee of the night, until I give him a treat. I can’t just say the word “treaty” because he knows I might not give it him. It’s happened ok? I’m not proud of it, but there have been occasions where I said “treaty” but didn’t actually deliver a treat. So he has learned that could happen, and now, he waits until I have it in my hand and throw it into the kitchen and then he comes in, let’s me wipe his paws and then gets the treaty from the kitchen. I’ve tried teaching him to wipe his own paws, but honestly, if I cant get my 24 year-old University student son to do it, I’m not sure why I have such high expectations of Bob.
Bob has obviously thought this through. His actions are deliberate and calculated to get the outcome he wants. He has shown some creativity in his solution. He knows I want him to come in. I want to go to bed. He stands just out of reach, on purpose, and if I try to grab him he’ll bugger off to the other end of the garden and then only comes back, tail wagging, when I’m back indoors.
There is no question in my mind that Bob is a sensient, loving, thinking, disobedient and creative Being who has his own ideas. I have no hesitation behaving toward Bob as I would anyone else. When I want him to give me something, I say “please”. When he gives it, I say “thanks”. Anything else would be rude and disrespectful. In every sense of the word, he is a member of the family and is treated accordingly.
We, and about 32 million other dog owners around the world, grant Bob and the canine masses this respect, despite the fact that Bob can’t talk and shows little understanding of any academic subject or the universe. I am prepared to bet good money that there are no circumstances where Bob could pass the Turing Test yet we still offer the respect and courtesy we might show another human being.
“The Turing Test?” you enquire. Yes, the Turing Test. Now, I know all of you know what the Turing Test is, but for the people reading who have been buried under a rock since WW2 (I’m not explaining that abbreviation, I’m really not), a brief explanation is needed.
This test is named after Alan Turing, the chap that figured out how to crack the Nazi Enigma Machine code during World War 2 (dagnamit!). Ze Germans used this machine to communicate real time troop, air and naval movement orders, so pretty important really. It was believed to be uncrackable because the encryption key was changed every day which meant you had to crack the key every day. Turing realised that every order was signed off “Heil Hitler” and this therefore provided a constant he could use to crack the entire code every morning. Basically he won the war for the allies. For his efforts, and in true British enlightened fashion, he was thanked by HM Gov. by being sent to jail for being a homosexual. Hurrah! Gawd blesss the British Empire Ma’am!
Anyway, the basic premise of the test is that if an interrogator (C opposite), is unable to tell the difference between responses from A or B (where A is a computer and B is a human), then the computer is judged to have Artificial Intelligence.
How long does it take you to figure out that you are talking to a computer when you have one of those automated Service Agents answer the phone? 5 seconds? Maybe 3? Imagine if you genuinely couldn’t tell the difference? Imagine if Freddie in Mumbai was actually an AI? Bad example, we all tend to get a bit irritated with Freddie, so lets go with Marge in Newcastle. You’d be polite and courteous to her because to all intents and purposes you would be talking to another being, and one with a natty Geordie accent, which apparently we all love. The fact that she is based on silicon wouldn’t matter because you wouldn’t know. If you were rude to her on the phone you would probably hang up feeling a bit guilty.
Well, an AI cant do any of that yet. Far from it. AI is very good at learning how to do narrow tasking but not so good in the arena known as General AI. Narrow AI is good though. So good, that it often surpasses human capability in the same task, very quickly. For example in March 2016 Google’s AlphaGo was able to beat 18 time World Champion, and 9 Dan (=very good) Chinese Go player Lee Sedol, 4-1 on its first outing. Go is the most complex game ever devised by man (or woman), with more possible positions available than there are atoms in the universe! So that’s pretty complex then.
But there are loads more out there and still more in the pipeline. In the field of Medicine, we now have AIs that surpass human ability to identify tumors from radiology images. Everything we know about autonomous driving suggests that it will be significantly safer than humans. Intelligent homes that combine personal assistants (Google Home or Alexa) with smart home devices (thermostats, lighting, camera’s etc) are already here. They are in financial markets, Business Intelligence software, travel experience apps, music streaming, shopping and, and, and, and. Narrow AI is already with us.
But no matter how “intelligent” these narrowly focused AI are, no matter how much data they can digest or how much machine learning they do, they cannot yet do, what Bob does, and they come nowhere near what Lu can do. Narrow AIs do not have ideas. They lack empathy. Their creativity is limited. They do not demonstrate understanding of the world, or the universe or of self.
But, they are getting there.
When General Artificial Intelligence (GAI) is achieved, how we treat GAIs and what rights we grant them, will be the issue of our time. We can treat them like we have every other minority in human history or we can treat them with “appropriate respect“.
Until then, I want Lu to be well mannered when talking to Siri, because I believe that if she finds it easy to be rude to Siri, she will find it easy to be rude to people. What does it say about us if we can’t be polite? How much easier will it be to integrate with future GAIs, if we have been treating their less bright forebears with respect and dignity? I’m not sure we will want them irritated with us, especially when they might be so much brighter than us.
