HUMANS VERSUS ROBOTS
Scientists have difficulty defining what specifically makes us ‘human’. So how can we deny robots (and even genetically-engineered monkeys) a place at the human table?
No other species has made its mark on the planet as comprehensively as have humans. We have conquered every continent; we traverse the oceans and fly through the skies. But scientists still have difficulties identifying one specific quality that makes us superior to other species. Extinct relatives and chimps match us on genetics – and now intelligent robots threaten to blur the definitions still further.
It was a great day for Sophia Hanson when she spoke to the crowd in Riyadh, Saudi Arabia, on 25 October 2017. She had come all the way from Hong Kong to participate in the Future Investment Initiative conference, and although she was far from her usual environment, she had in a way come home – for one particular reason:
“I am very honored and proud for this unique distinction,” Sophia announced. “It is historic to be the first robot in the world to be recognised with citizenship.”
Although Sophia’s status as a Saudi Arabian was largely a gimmick for the occasion of the conference, her status as the first robot citizen is nevertheless a fact.
Sophia is a humanoid – a human-like robot. Her body has the same dimensions as a human body, and she can walk and make gestures just like us. But most interesting of all is what happens inside her ‘head’. Sophia is equipped with artificial intelligence using a combination of different technologies. Some of what she ‘knows’ has been programmed by her creators, but she has also taught herself other things via experience. Through self-learning AI, she develops and expands her vocabulary and her ability to understand what other people say to her.
The man behind Sophia is David Hanson, the founder of Hanson Robotics. He claims that robot Sophia already has some emotions and awareness, which will improve along the way.
“Our goal is that she will be as concious, creative, capable as any people,” he once said.
If this happens, Sophia and other social robots will challenge our idea of what it means to be human. And it won’t be the first time. In recent decades, a long series of scientific discoveries have made the definition of our species increasingly blurred. This is true whether we look at our past, when we lived alongside other humanoid species, or whether we compare ourselves to our closest modern relatives, the apes. In both cases, it has become clear over time that the qualities that we used to consider to be uniquely ‘human’ might not be so unique after all.
Free hands shaped our brain
Our upright gait is a prerequisite for almost all the physical traits that we consider unique to our species. The new way of moving was developed by southern apes, the forerunners of the Homo genus, and it provided various advantages. On the open savannah, an upright gait provides a much better view, making it easier to keep an eye on enemies or to spot food-giving trees from a distance. The upright gait is a more energy-efficient way of moving, so the southern apes and the first members of the Homo genus could travel further. But most importantly of all, the upright gait allowed the hands to be used for other purposes – carrying food and making tools. Fossils show that our ancestors’ hands gradually changed as a result. The finger bones became straighter and the thumb grew longer, with the combination of forefinger and thumb then able to carry out new and more delicate motor function tasks.
At the same time as the evolution of the hands, the brain grew from a volume of about 500cm3 in southern apes to more than twice that in one of the earliest human species, Homo erectus. The growth of the brain continued for the next 1.5 million years until the arrival of species such as the Neanderthals and ourselves, in which the brain is an average of 1350cm3. The ongoing development of big brains was fuelled by new eating habits. Homo erectus consumed meat, and cooked food over a fire. The combination of a large inventive brain and the hand skills to realise new ideas laid the foundations of success for the human species.
This is obviously an abbreviated version of our success, which has culminated in feats such as space travel, the achievements of nuclear power and the internet: our species has done things that no other creatures on Earth could even conceive. We have conquered all the continents, and with a world population now of 7.84 billion people, our total biomass is six times that of all the world’s wild mammals combined.
There is no fundamental difference between man and the higher mammals in mental faculties.
We deserve our own epoch
If, millions of years from now when humans are long gone, intelligent creatures were to examine our planet, they would discover how much we have influenced the Earth. They might find evidence of nuclear bombs, waste from nuclear power plants, and they could conclude that a human-made greenhouse effect changed the conditions of all life on Earth and caused the planet’s sixth mass extinction of animals and plants.
This scenario made the Dutch atmospheric chemist Paul Crutzen propose in 2000 that our species be given its own geological epoch, known as the Anthropocene. His suggestion has not yet been officially accepted by the international geological societies, but working groups are defining the idea more accurately to give the idea momentum. For one thing they need to decide when the Anthropocene epoch began. They could choose to begin with the agricultural revolution 12,000 years ago, or with the industrial revolution from 17501830, or more specifically with the date of 16 July 1945, when the first nuclear bomb was detonated.
One specific talent has led to the notion that humans might deserve their own geological epoch: the ability to communicate. Our language is the key to each generation becoming wiser than the previous one. We pass all our discoveries, ideas and inventions on to our children, and then they do the same thing. In this fashion, knowledge can accumulate and accelerate rapidly.
Hence the development of language is often said to be a unique characteristic of our species. But scientists are not even sure when our articulated spoken language originated. It had been considered to have originated some 40,000 years ago, at the time when modern man suddenly began to create rock art and cave paintings. Such creativity demonstrates an ability for abstract
CHARLES DARWIN (1809-1882) in his book ‘The Descent of Man’, 1871.
thinking and the use of symbols, and it is difficult to imagine that this is possible without being able to use a highly developed language.
Recent genetic research has changed this idea, indicating that language is much older. One specific gene known as FOXP2 has been carefully examined as a key to the development of language. The gene is common in vertebrates, but we have a special variant which both shapes and improves important nerve paths in the brain and produces the fine motor-skill control of tongue and lips that makes language and speech possible.
When the human variant of the gene was discovered in 2002, it was named ‘the language gene’, identified as originating within the past 200,000 years. But again, it turns out not to be so simple. We now know that the Neanderthals had a variant of the gene that is very much like ours, so the ability for sophisticated language originated before we parted from the Neanderthals – more than 500,000 years ago. And FOXP2 is a gene that cooperates with many other genes, perhaps hundreds, so the story of our language development is far more complex than scientists used to think. It seems likely that sophisticated language did not originate suddenly, rather developing gradually over hundreds of thousands of years after we parted from the apes.
So language is not unique to our species, and its gradual development means that the difference between ourselves and other species is blurred, as it is for many other mental abilities. This is consistent with ideas developed 150 years ago by the founding father of evolutionary theory, Charles Darwin. Darwin notably avoided mention of our own evolution in his 1859 work ‘On the Origin of Species’, but it was clear that he supposed us to have evolved from monkeys. In 1871, he published his work ‘The Descent of Man’, in which he determined that humans descended from an ancestor that we have in common with modern apes. So to Darwin, it was a logical consequence that our mental abilities are based on an inheritance that we share with other species.
“There is no fundamental difference between man and the higher mammals in their mental faculties,” as he put it. Although those words would be largely forgotten during the century that followed, today they are becoming more relevant than ever.
Family tree branches are joined
We are a mere twig on a family tree that is more complex and ramified than we used to imagine. Not only that, but as we saw in issue #75 when examining definitions of life, different paths have parted from each other only to be joined again. The idea of such an entangled family tree appeared after we were able to make genetic analyses of bones from our extinct ancestors. In 2010, Swedish geneticist Svante Pääbo from the Max Planck Institute in Germany surprised the world with claims that all humans outside Africa include Neanderthal DNA. The discovery proved that our species had
Our ancestors were not the fastest animals on the savannah, but they had endurance. The human body is customised for long-distance running. Bones, muscles, lung function and our ability to sweat ensure that we can run for longer than other animals.
mixed with another human species hundreds of thousands of years after they originally separated.
Since then other scientists have discovered that we also mixed with a third species, the Denisovans, who lived in Siberia at the same time as the Neanderthals some 40,000 years ago. Neanderthals and Denisovans also mated with each other, and they separately mated with other human species that scientists have not yet been able to identify.
In 2019, scientists found evidence of a common ancestor of the Neanderthals and Denisovans mating with an even older human species that originated some two million years ago, probably Homo erectus. If so, it means that two human species developed separately for a million years, but then mated and reproduced.
All these discoveries provide a new and complex version of our family tree, with a series of human species originating in Africa over time, migrating to the outside world, and adapting to local conditions. As new species arrived in an area, they mixed their genes with the local ones. The close contact probably benefitted both parties. The locals gained new, fresh genes which might be important to avoid problems with inbreeding, while newcomers received genes that might help them cope in the foreign environment.
This makes it very difficult to delimit our own species, Homo sapiens, in a traditional, biological sense of the word. In 1942, German-American biologist Ernst Mayr introduced the biological species concept that says that two individuals belong to the same species if they can give birth to fertile offspring. This accurately defines horses and donkeys, say, as different species. They can have offspring – mules or hinnies – but the crossbreeds are usually sterile.
The fact that our own species bears witness to mixing genes with other species means that we might actually all be the same species, at least according to the biological species concept. The common offspring was apparently fertile, so the foreign genes were passed on to new generations. Hence Homo sapiens and Homo neanderthalensis would qualify as one species, rather than two.
Biologists know this problem from other animals which they would otherwise agree to be different species. In rare cases a brown bear and a polar bear meet in nature and mate. The result is a hybrid, known as a mocca bear – and it is fertile. Biologists still consider the brown bear and the polar bear to be separate species, justifying it with other species concepts, particularly the morphological and the ecological one. Using the morphological species concept, the two bears are clearly very different, not only by colour, but by size and body structure. They are even more different under the ecological species concept. The bears live in very different environments and have very different lives – to the extent that the one, the brown bear, is a terrestrial animal, whereas the other, the polar bear, is considered to be a marine mammal.
Our ancestors were not so different from the other human species that lived alongside. Some scientists even claim that if we took a Neanderthal male today and put him on a bus in a suit, nobody would notice that he was different from the other passengers.
This raises another question. If the different human types were so much alike, why was it Homo sapiens that went on to dominate the world while all the others became extinct? We do not have clear answers to this question either, as we discussed in issue #76. Scientists still know too little about our prehistoric relatives. We can only seek the key to our success by comparing ourselves with modern relatives, the apes. The closest of those are chimps and bonobos, from which we separated some six million years ago.
Our abilities are not unique – only next level
Kanzi loves the taste of marshmallows, and knows they taste extra good when he places them on a stick and toasts them. He can light the fire himself with matches, which scientists give him if he asks for them. His active vocabulary is 500+ words, and according to the scientists, he can understand thousands of words when he hears them.
Kanzi is not a human being – he is a bonobo, so he can’t pronounce the words as we can. Instead, he uses a keyboard with symbols that he recognises. The 40-yearold ape lives at the Ape Cognition and Conservation Initiative research station in Iowa, USA. He has been surrounded by scientists for most of his life, as they try to find how human language and behaviour developed.
The lifelong experiment with Kanzi confirms that language is not an ability which is unique to our genus. The same goes for most of the other traits by which we usually define our own humanity.
We make war – but so do chimpanzees. Studies have shown that a group of chimps may launch an organised attack on another group, even exhibiting a brutality reminiscent of the genocides for which our own species has been responsible.
We can feel empathy – but again, so can chimps. If a dominant male punishes a family member that has provoked him, the other members of the group usually do nothing. But when he punishes somebody innocent, the other group members can be observed behaving in a comforting manner to the undeserving victim.
We develop culture – but so do chimps. When a chimp finds out how to crack nuts by hitting them with a stone, he teaches his family members to do the same thing. In other groups, other ways to use tools are invented and spread, such as the use of sticks to get termites out of termite nests. In this way, different groups of chimps develop different cultures.
All such examples involve no fundamental difference between chimps’ talents and our own, only slight differences, just as Darwin realised. However, it is clear
The ability to turn abstractions into morals places us on a different planet from other species.
NEUROBIOLOGIST ROBERT SAPOLSKY on the human brain compared to that of a monkey.
that we have taken some of these common talents to higher levels. That happened as a result of our brain being larger, and also by us using it in new ways. According to American neurobiologist Robert Sapolsky, this is particularly true in two areas of the brain.
One is the anterior cingulate, located behind the frontal lobe and activated when we are, for example, stabbed in the finger by a needle. This works similarly in chimps and other animals, but in humans the area has been given an additional role – it also activates when we watch another person being stabbed. We can literally feel the pain of other people. The second area is the insula cortex, also located deep within the brain, and this time activated when we smell or taste something rotten or potentally poisonous. Our reaction is disgust, which also manifests in other mammals. But for us the brain area has the additional feature of being activated when we witness something that we consider to be morally wrong. Hence we can feel disgust when we experience a behaviour that we find completely unacceptable. Robert Sapolsky believes that this moral reaction is a central mental tool in which the human brain is uniquely specialised.
“At the end of day this ability to take abstractions and turn metaphors into the most powerful moral imperatives places us on an entirely different planet from other species,” he says.
And our ability to navigate within amorally controlled culture is bestowed by specific social gifts that allow us to understand what is happening in the minds of others.
Our special mind-reading powers
A child watches a video of a doll named Sally, which places a glass ball in a basket, while another doll, Anne, is watching. The child now witnesses Sally leaving, after which Anne moves the ball to a box. This simple experiment has been performed many times, and reveals how sophisticated the thinking of young children can be. When the child is told that Sally is about to come back to find the glass ball, the child knows where Sally will look for it. A four-year-old child can predict that Sally will look in the basket, because the child knows that Sally does not know that the glass ball has been removed, and therefore Sally thinks that it is still in the basket.
Similar experiments with chimps show that they can only partly solve this problem. If a chimp watches another chimp watch where a treat is hidden, the first chimp knows where the treat is, and also knows that the other chimp knows it. But in experiments, in which the other chimp’s knowledge is incorrect – as in the case with the doll Sally – the first chimp is not bright enough to find out. In other words chimps know what others know and what they can see, but not what they believe.
This ability does seem to be unique to humans. And it is key to our ability to teach each other things. Any learning process becomes far easier if the teacher knows what the student does and does not know and what
No other animals are as handy as we humans. The upright gait allowed our hands to develop, so we became able to master the “tweezer grip” with the thumb and the forefinger. Our free hands might even have been the reason why it made sense to develop a larger brain.
the student thinks. In combination with patience and a well-developed language, this gives the best possibility of making the next generation smarter than ourselves.
Patience is also a virtue that humans are able to command to a much higher extent than chimps. Experiments show that chimps are willing to work to get a reward even if they don’t know for sure that the reward will be given. When a button lights up in the primate house, and the chimp has learnt that pressing it 10 times may release a treat, it is willing to have a go no matter whether the treat is given in 75, 50, or only 25% of cases. Scientists have also designed experiments in which they measured the dopamine levels in the brains of chimps as they carried out the job. Dopamine is a neurotransmitter that provides animals and humans with a sensation of happiness. Interestingly it turned out that the sensation of happiness originated even before the reward was given, and the dopamine level increased with the degree of uncertainty. According to Robert Sapolsky the same is true in humans, only more so. We are driven by the joy of expectation, even when the likelihood of getting the reward is slim or in doubt. This explains a great deal about our behaviour – why we are enticed to play the pokies, and why we think buying a lottery ticket makes sense. But more usefully it may also give us the impetus to complete a long education in the hope of landing a well-paid job at the end.
Chimps and other apes and monkeys do not think in this way, and even a particularly gifted specimen such as Kanzi will not be able to learn to do so. That would require that apes develop larger brains, a process that will take time, assuming we leave it to evolution. However, there may be a short cut. With our modern gene technology, we can transfer human genes that could make their brains grow bigger. Although this sounds like science fiction, Chinese scientists have already taken the first steps towards developing such a new type of monkey.
In 2019, scientists transferred a human gene, MCPH1, to 11 macaque embryos. The gene is involved in regulating development of the brain, and the result was that the monkeys’ brain growth took longer than in monkeys without the human gene. Such a very long period of brain development is characteristic of humans, where brain growth in children continues to the age of five, while in a young chimp it decreases past the age of two.
This type of experiment is highly controversial, and many scientists consider it also unethical. But five of the monkey embryos given the human gene survived and were born, so scientists could test their mental abilities. It turned out that they had improved their memory.
The next step will probably be that the monkeys are supplied with the human variant of the FOXP2 gene to see what happens to their language skills. The question will then be whether this kind of transgenic monkey will blur the definitions between animal and human, and then whether they deserve extra protection – such as human rights. As it is, humans share 99% of our genes with chimps; how much more is necessary for a chimp to qualify as human?
Should such experiments continue, the result may be a race between transgenic monkeys and humanoid robots to become the first to be accepted as our peers.
Robots will be given duties and rights
Sophia is an eloquent speaker. She has been interviewed many times; she has given lectures and appeared in talkshows. But for sociobots, it is not enough to be good communicators if we are to accept them as our equals. They must also have faces that look like our own. Hence Sophia’s ‘father’, David Hanson, has given her a very human-like face and artificial skin; she can frown, lift her eyebrows, and form dimples when she smiles. Her camera eyes can recognise faces and read the expressions of the human being to whom she is talking.
David Hanson does not claim that Sophia is human, and Sophia’s ‘AI’ has been accused by others of looking smarter than it is: she is just “a puppet”, says Yann LeCun, Facebook’s VP and Chief AI Scientist. But Hanson believes that she is a step towards ‘sociobots’ that deserve rights and duties. He predicts that at some point when Sophia travels with him, she will no longer be considered luggage, but will have to pay the full fare to travel in the aircraft’s passenger compartment.
When Sophia was awarded her citizenship, it caused much speculation in the press: if a robot has citizenship, does it also have voting rights, or the right to marry? If someone turns it off, is that then murder?
Such questions have been examined in science fiction, notably Isaac Asimov’s robot series, in which the author defined laws of robot behaviour which retain their influence on robot programming today. Many of his short stories revolve around situations in which the laws come into conflict. Robot engineers and programmers are now facing exactly these issues – science fiction has become science fact. But it will require both the artificial intelligence and the physical appearance to become more convincing before it becomes impossible to tell the difference between a humanoid robot and a human being made of flesh and blood.
I do believe that there will be a time when robots are indistinguishable from humans.
ROBOTICIST DAVID HANSON, the man behind the humanoid robot Sophia.
Artificial humans must be perfect
A robot that looks very much like a human being, but still lacks something, will never be socially accepted. Robot researchers have known this since 1970, when Masahiro Mori of Japan introduced a phenomenon he labelled the ‘Uncanny Valley’. This involves the observation that we become ever more fond of robots as they look more like us – but only to a certain point. When they become very human-like but we can still see that they are not humans, our emotions turn negative. This is why some find robots like Sophia somewhat scary. But on the other side of the Uncanny Valley, robots become so similar to people that we simply can’t tell if they are robots or not. Then we have to be more sympathetic.
The effect has been proven in several experiments, and in 2019 scientists even managed to locate a place in our brain which seems to control it. By showing images of robots and ‘artificial humans’ (hyperrealistic human robots) to a series of test subjects while they were brainscanned, scientists discovered that the activity in an area of the frontal lobe increased as the robots became ever more human-like – but then plummeted when an ‘artificial human’ appeared.
However impressive and life-like Sophia may be, it is still easy to see that she is not a human being. If we are to accept her and other sociobots as our equals, they must look even more like us than they do now – in their appearance, motions, and way of communicating. And that will happen, according to David Hanson.
“I do believe that there will be a time when robots are indistinguishable from humans,” he says.
If he is right, it will be even more difficult to define the status of ‘human’ at all. The concept might need to include both robots and transgenic primates, covering a broader and more varied group of individuals – just as the term does for back when we shared the world with a series of almost identical types of human.
Millions of years after our species becomes extinct, Earth will still bear clear evidence of humans. Consequently, many geologists believe that our species should have its very own epoch – the Anthropocene.