The Ethics of Boosting Animals from Sentience to Self-Aware Consciousness
The question of boosting the intelligence of animals other than humans, to the level of self-awareness and consciousness, is one that cannot be considered in isolation from the general process at work in the social and speculative movement known as Transhumanism. Transhumanism is partly about augmentation of the human body, for example by molecular computational nanotechnology (cell repair and DNA repair molecular machines, and rod logic molecular supercomputers for enhanced memory). More generally is is the movement to ‘self-directed’ evolution (replacing evolution by random mutation and natural selection). This was pointed out by myself in 1996 and by Damien Broderick in his book The Spike, which was the first popular book on the idea of the ‘technological Singularity’ (a time when technology surges forward based on accelerating computing).
This is a radical movement away from evolution as we understand it, and unchosen death. It is a movement towards chosen lifespans (and very extended ones, certainly some claim possibilities of hundreds of millions of years) and augmentation.
In this scenario, it is the internal environment that can alter the physical characteristics of a population. Make a mistake by modifying something essential to your will to live and you may well end up permanently dead.
For a conservative Transhumanist, the minimal rational position would be to consider us already effectively immortal. This is because genetic and proteomic therapies, along with cures for many diseases, will begin to be available within a very few years, perhaps the middle of the decade of the 2020s.
This will tide most people over until the arrival of computational superintelligence and computational nanotechnology (that is, not materials nanotechnology, but ‘intelligent’ nanotechnology, devices at the molecular scale that cannot reproduce but that can be ‘commanded’ to achieve goals). Stem cell therapies will provide a surface appearance of youth in terms of hair and wrinkles and teeth (this is already happening, in the case of teeth see the work of Dr Mao of Columbia University). The fuller issues of a long life are more complex (memory, consciousness, cosmic ray damage, skeletal structure and so on). However, telomerase therapy (for DNA repair), cell-sized ‘nanobots’ and smaller nanotechnological cell repair machines and molecular supercomputers (perhaps sixty in each cell) should be able to monitor and augment of both individual neurons and a person’s wider neural structure.
As a comparison to these ideas of living for many millions of years, the oldest living multicellular thing on Planet Earth currently is the clonal root system of the quaking aspen Populus tremuloides in Utah, dated to 80,000 years, probably older than modern humanity itself. In theory some of the ‘immortal’ jellyfish, Turritopsis nutricula, could live forever but in practice most are probably killed in the environment. Many animal species live longer than humans, for example the rougheye rockfish lives for 205 years.
It is important to understand that superintelligence will be required to solve deeper problems of memory and neural identity over the immense time scales projected, especially for those conservatives who choose to remain relatively human for much or all of the time.
Some radicals may gradually move their neural structure into AI systems, or even into virtual realities within computer substrates (for example topological quantum computers or femtocomputers, both of which are devices which have not been built yet).
It is hard to argue against the physical possibility that such real human-AI or virtual beings could live until the end of the universe. For example, Abraham Loeb of the Centre for Astrophysics at Harvard University believes intelligences could orbit around red dwarfs for ten thousand billion years.
This then is a general structure for the consideration of ‘boosting’ of non-human animals.
There are a number of obvious issues. I will leave aside, for the moment, the issue of what model should be used in ‘adding’ to the brain of a non-human animal (rather than just ‘enhancing’ such a brain).
What would be the extent of such ‘boosting’?
Generally there are wild animals, domesticated livestock and domesticated pets. Domesticated livestock number roughly half the human population and occupy 25 per cent of ice free land. Domesticated and other pets number several hundred million.
There is also a potential division in terms of levels of intelligence, with anthropoids such as apes and chimps, dogs, cetaceans, octopuses, pigs, parrots, crows and so forth at the more aware end, often equivalent to two to three year old humans and capable of possessing learnt language and a miniscule degree of self-recognition.
At the bottom of the animal kingdom is the only creature whose brain has been fully mapped, the humble 302 neuron, 9000 synapse, millimetre-long, dirt-dwelling worm, Caenorhabditis elegans.
Full boosting of all let us say all animals above 10 million neurons would encompass mice and above.
If the biosphere was artificially separated or burst in such a fashion there would also be a range of other issues, for example the issue of maintenance of the plant kingdom and the issue of carnivores. In regard to carnivores (and many humans are carnivorous) this problem could be solved by in vitro or artificially grown meat. Many groups are now working on this, including NASA, partly because it may be more efficient for food production, and animal welfare groups such as PETA have argued for it.
Then there would be the issue of addressing any emerging or evolving animal species as they move to the threshold level, whatever that might be.
What is the general background to the emergence of human intelligence and why would we feel morally obliged to help other animal species, species that many persons would label as simply empty shells, and therefore beings that while alive have nothing to boost?
Each of the five extinction events (leaving aside the current and largely unconscious or unvoted-for one that we have induced) that the planet has faced since the emergence of multicellular life has killed most things larger than the size of a cat and each time in this period of 500 million years multicellular life has bounced back. Effective death from changes in the Sun, in the absence of a technology capable of moving the Earth in its orbit (now possible with asteroid herding) would result in the extinction of all life on Earth in perhaps around 500 million years time from now, maximum (though estimates vary).
Thus multicellular life on Earth was perhaps half-way down the track when it won the intelligence lottery.
Further natural extinction events would also be somewhat likely despite a lessening in solar system collisions over time.
Over the last thirty million years the primate order produced several species with a spoken language: Homo ergaster, Homo heidelbergensis, Homo neanderthalensis, Homo denisova, another as-yet unnamed species (see New Scientist of 30 July, ‘Our Hybrid Origins’ page 38), Homo floresiensis and Homo sapiens. Our species, as the only survivor (though we incorporate 15% of our DNA from neanderthal, denisovan and the unnammed species) has inherited the responsibility of ultimate moral power.
Not only can we direct our intelligence and mortality, for ourselves and artificial intelligence, but we will be able to direct nature, at least as it applies to life. “With great power comes great responsibility.”
What would be the arguments used in favour of such a radical realignment of life, with immortal (by choice) AIs, humans, hybrids and animals?
Secondarily, is it legitimate to use the Homo sapiens, mammalian template as a model for creating self-aware consciousness, culture and language?
In addressing these questions I take the brain as the seat of consciousness, and ignore any religious belief in a ‘soul’.
The first argument is reciprocity. Our example is important in terms of teaching the first AI or AIs while it or they are very briefly children.
If we expect a superintelligent AI to provide us with timely provision of the mechanisms of superintelligence, immortality and perhaps even the theory of everything of physics, then we must have lead by example, even if it is regarding something which we have only been able to support in principle, while lacking the immediate means to achieve it.
We may already possess what might be termed ‘superintelligence of the imagination’ – as a science fiction writer I have read much of this amazing breadth of thought. Amongst my earliest reading material was the famous 1944 novel by Olaf Stapledon, Sirius, concerning a dog whose intelligence was boosted. However, as we can all attest, outside of our imaginations we have limitations of the mundane.
The most everyday languages we can learn is about sixty, the most faces we can remember (an this is one of our key strengths) is about ten thousand. Only about thirty thousand logic trees go into each thought or action.
Even with advanced computational nanomachines embedded in our cerebrospinal fluid – according to TIME magazine, perhaps the equivalent of two hundred thousand human brains – we would still require further superintelligence to make a workable and permanent system quickly. In other words, if we expect to fix ourselves up within a century or less, we will need the assistance of Artificial Intelligence.
The second argument is empathy. While many of us may not feel we have much empathy with a leech, quite a few have empathy with social animals and pets, and indeed love for them.
The third argument is responsibility. If diversity and potential diversity is a positive feature of mortal life, one we now look for in our children as reflections of ourselves, then it should be reflected in immortal life.
Unlike Peter Singer, the popularizer of Animal Liberation, I believe that continuity is important. That a human baby is no different from a dog for some months is no grounds to ignore its future state. Similarly, with the potential abilities of the near future, we could regard a dog as a gestating ‘super-dog’.
The fourth argument is that it is fun to have more immortal beings around, with which we can interact.
The fifth argument is that it is fun to have more kids to raise.
The sixth argument is that under future conditions such as full automation, distributed production, individuated design, robotics, 3D printers and eventually nanotechnological assemblers (a much more advanced form of 3D printers, also known as nanofactories or molecular manufacturing) it will not cost us anything.
While I am not advocating megascale solar sytem engineering I also note that 1990s estimates of the maximum number of beings capable of being supported by all the material of the solar system as about 600 billion and certainly there is a lot of scope for all the current inhabitants of the planetary biosphere.
The seventh and final argument is that by making a current adult human the minimum requirement for a neural map for a birthed consciousness (along with the normal developmental period of adolescence) we have a reasonable sliding scale that involves no moral guilt for failed help.
I consider it possible that even superintelligences might indulge in ‘stepping down’ to a current human mode of complexity for a while, as an indulgence or holiday or for the many simplicities of our style of life!
Now to the question of why it would be alright to use the human neural template as a model for what is not present in a developed sense or at all in other animal species. Such a template would not imply a necessity to humanize form (such as hands) but would relate to issues of self-awareness (such as the ventromedial prefontal cortex), the ability to conceptualize manipulation of the environment, curiosity, learning (such as the dentate gyrus of the hippocampus), language and, to the necessary extent, co-operation.
The process of boosting is one of addition, not of subtraction. The fact that what is added may have affinities with human biosociology, rather than being extrapolations of perceived species traits, is not relevant. That is part of the cultural baggage of Planet Earth and entirely appropriate. A baby does not consent to birth, but its parents do. In a sense, we operate as the collective parents of the animal world, in a developmental sense, in addition to the directly biological parents of each animal.
If this process of a technological escalation or Singularity has occured on other worlds, as I suspect it has, I believe the logic of necessity would have produced similar situations elsewhere.
That is, those worlds in which creatures like octopuses manipulate their environment and construct computing devices end up with octopus templates for other species. This is not to say that voluntary divergence of immortal form may occur after technological escalations or Singularities but a template is a type of necessary minimum for emergence into controlled immortality.
I would also emphasize that in my view consideration of a neural map or neural networks by themselves are not the complete story. Our memetic map (at some levels a cultural map), though less easily detectable, is just as physical as DNA.
Of course, the hardwiring of the brain is the initial basis of our every thought and emotional state. But the conceptual framework, the map that produces the great thought, for example an historical insight, is not just the result of one individual. It is a matrix of billions of events, eddies of acts we have never seen directly, some of which disappear but most of which flow over us culturally, the accumulated decisions of many many generations.
This background produces enhanced intelligence in a very direct way. Without it, growing up marooned on a hypothetical friendly desert island, we would be intellectually impaired at nearly every level by the end of childhood.
In a sense, the impulses behind the Stonehenge-Woodhenge complex are the same impulses producing the technological escalation or Singularity.
I also believe there are many more types of intelligence than some of the categories that I have read about in scientific journals, and that intelligence, like history, is at many levels an art form.
There are plenty of people who are brilliant at cryptic crosswords and idiots as painters or politicians. Intelligence, in my view, is often for example as much about awareness as problem solving, or as much about successful mapping as paradigm shifting or analogous logic.
Scientists have a natural tendency to want to look to intelligence that is capable of being easily verified in experiments. Of course, all types of intelligence have mechanical, neurological underpinnings too, but that should not move us away from the observation that the same platform can have different characteristics, that for example an intelligent aesthetic can flow from emotional understandings that are not simply elegant equations but matters of deep and permanently grey shadings. In other words, a car can behave utterly differently when driven by another driver.
To those who would say that this very discussion is unnatural, I would argue the opposite. This is the natural culmination of a natural process. Self-directed evolution and the boosting of ourselves and other animals (and in fact AI itself, were it to be deliberately curtailed in some fashion) is the logical and inevitable extension of previous, evolutionary processes.
This is the same logic that tells us that superintelligences are likely to have more emotional states and more empathy than the average mortal person of the past, not less.
Though I find the human mind a wonderful thing. It largely refuses suicide despite the apparent limitations of arbitrary inheritances of suffering and inevitable death. The majesty of this resistance is what is known as ‘the human condition’.
Humans remains largely positive and altruistic socially, despite several per cent of males (and far fewer females) suffering from empathetic underdevelopment due to sociopathology.
I understand the goals of the social movement of Transhumanism, which began in the 1990s, as a synthesis, as a fusion or set of multiple paradigm shifts. In a sense it is the ultimate extension of the 1970s scientific theory of planetary operations popularized by James Lovelock and known as ‘Gaian theory’.
This fusion is not a threat to the natural world or our cultural tradition, but the best hope for a diverse and universally aware community, with a voluntary immortality. The technological Singularity can embrace a universal democracy of sentiency and all beings on planet Earth.
Avatar Polymorph (no surname; avatar as in virtual) is a writer of fiction (The Prisoner Gains a Blurred Skin, Black Pepper Publishing) and science fiction and a former policy adviser to Australian heads of government. Born in 1961 on the day that Alan Shepard was launched into space, he writes about futurology, ethics and spirituality.