Improving humanity’s survival chances in a future world run by robots is a two-step process.
Step 1: Be kind to animals.
Step 2: Let robots learn from us, hopefully replicating our kindness to animals in their behaviour to humans.
By George ILIEV
We should treat animals kindly in the hope that robots will treat us kindly once they take over. History does not necessarily repeat itself but it rhymes, to paraphrase Mark Twain. Or you may recall the story of Roman general Scipio Aemilianus who wept when he destroyed Carthage in 146 BC, as he foresaw that the same fate would strike Rome one day. If we treat animals as enemies, we shouldn’t be surprised if we get similar treatment in the future.
1. Pattern recognition: monkey see, monkey do
We are already observing AI (artificial intelligence) developing racist and misogynistic tendencies based on analysis of people-to-people communication and human culture. It is likely that the robots of the future will take their cues from our behaviour, as humans will provide most of the data on which robots will train their AI.
Historically the treatment of animals by humans has run along a gamut of mostly negative experiences: hunting; raising animals for slaughter; domestication to use as beasts of burden; and experiments on animals. Yet, recent history also shows signs of hope: the modern trends of taking care of pets with kindness and respect; or giving up animal-derived food and becoming vegetarian/vegan.
Our treatment of animals is a trend-setting pattern that robots will likely draw upon or even benchmark against when they fine-tune their policies and behaviour towards us.
2. Conformism is the old normal (Watch Zootopia!)
Our peaceful coexistence with robots rests on the assumption that robots will copy us when we are on our best behaviour. They may or may not do so (for better or worse) but either way we should strive to provide positive models to copy rather than ones that may harm humanity.
We can see in the evolution of animal culture and social learning (learning from others) that animals are conformist creatures. Experiments on various species of monkeys and apes demonstrate this clearly:
“Animals strive to act like others, especially others whom they trust and feel close to. Conformist biases shape society by promoting the absorption of habits and knowledge accumulated by previous generations.”
(Frans de Waal, Are We Smart Enough to Know How Smart Animals Are?, P. 256-259)
If all creatures with a degree of consciousness are kind to one another, then robots will not find alternative models to observe and copy and will conform to our culture. In other words, we should be building a world modelled on peaceful coexistence with animals as in Zootopia (Zootropolis, the movie) before the robots have developed sufficiently to redesign the world on their own. If we build "Zootopia", all we would then need to do is ensure that robots have enough time to evolve around us and with us, so as to learn from our behaviour.
3. Predator vs Alien
The assumption that robots will copy our best selves may be invalidated if robots decided to copy patterns from human history (e.g. slavery) or existing predator-prey behaviour among animals. However, predators hunt their prey to survive, while the robots of the future are unlikely to need to eat flesh to charge up their batteries. This key metaphorical distinction is unlikely to be lost on machines that have evolved intelligence and consciousness.
The key question in the behavioural conformism hypothesis then is whether robots would feel as similar to us as we feel similar to other animals, or as alien to us as we feel alien to other animals. Humanity seems to be treading a fine line between the two at present. Would robots resist the metaphorical model of “predator” or "master", including holding back from playing with humans as an object of entertainment (e.g. as in fox hunting, or in the circus)? Or would they, while being aliens to us, evolve into benevolent aliens that take the best from humanity and disregard historical mistakes in our treatment of other human beings and other animals?
4. 100 Years of Biological Solitude
We will find out in the next 100 years or so which way the robotic cookie crumbles. Our solitude of living only with other biological species will be over when robots become able to reproduce. But if we are still cruel to animals when the first sentient robots come into being, we may run out of time to convince the next generations of robots that we ourselves are worth preserving.
