Animals ranging from parrots to elephants continue to challenge our perception of consciousness, long-held as a uniquely human trait. But the reaches of consciousness don’t stop at animals. As artificial intelligence gets smarter, we are faced with moral dilemmas of how machines could one day not just think but also feel.
The ethics of consciousness, not just in humans but also animals and machines, is complex. To try and make sense of it, research is currently underway to develop a method for objectively measuring consciousness — a formula that could explain how aware any living, or artificial, being is.
A HISTORY OF CONSCIOUSNESS
The concept of consciousness — our awareness of what we experience, and what those experiences mean — has long been debated. Descartes’ exploration of what it means to think — “I think, therefore I am” — was written in 1640, but our understanding of how the brain works is still limited.
Ned Block, professor of philosophy, psychology and neural science at New York University, identified the three major theories of consciousness prevalent in modern philosophical thought; the ‘higher order’ theory, the ‘global workspace’ theory and the ‘biological’ theory.
Means of measuring consciousness have prompted debates about how ethical the practice would beHuman Connectome Project
Higher order theory states that consciousness is the awareness of an experience; you perceive a sensation of ‘redness’, for example. But you also have a higher-order awareness of this, a perception of your perception — and that, according to this theory, is consciousness.
Global workspace theory posits that a large set of matched pairs of conscious and unconscious processes make up our consciousness, and the biological theory argues that consciousness is some sort of state of the brain.
That human beings experience consciousness isn’t under debate — but just how this consciousness results from tangible brain activity is. Despite many attempts to explain the link between experience and embodiment, the question of consciousness remains a problematic mystery.
MIND AND BRAIN
But now science is trying to solve the quandary. Giulio Tononi, a neuroscientist at the University of Wisconsin, is developing a theory that is attempting to empirically measure consciousness. Integrated Information Theory, or IIT, argues that consciousness could be present in any number of things — not just humans but also animals and even smartphones and computers. The caveat? The information contained by that structure must be adequately integrated and organised.
“We usually grant consciousness to others — of the same kind we experience in the privacy of our own mind — if they can tell us what they feel, or if they look and behave more or less like us”
Giulio Tononi, University of Wisconsin
Levels of consciousness are captured in a value Tononi refers to as Φ, the Greek letter phi, which is also the symbol for the mathematical golden ratio. In Tononi’s system, the value of phi represents the ability of a neural network to integrate and organise information. The higher a network’s ability to do so, the higher the phi — and the higher the level of consciousness. This is all captured by a device that electrically stimulates the brain. This device, according to Tononi, is able to measure just how integrated a network is.
Neural conditions in brain damaged patients can differ hugely from those in healthy subjectsShutterstock
“One of the promises of IIT is that we would have a quantitative measure of ‘level’ or ‘amount’ of consciousness,” Professor Ron Chrisley, a reader in philosophy from the centre for cognitive science at the University of Sussex, tells WIRED.
“And that would be a useful thing to have for a variety of reasons, some having to do with furthering our understanding of consciousness and some being more practical, such as evaluating the effects of various drugs, including but not limited to anaesthetics.”
There are a few limitations, though. A human brain can be conscious, as can the brain of a cat or dog — but larger structures, such as social groups or communities of living organisms (like forests) can’t. This is because every element of a structure (the neural cells of a human brain, for example) have to be sufficiently interconnected. Each cell in a human brain is linked to millions of others — and this, rather than a more broad concept of integration, is what indicates consciousness.
THE CONSCIOUSNESS OF CEPHALOPODS
IIT also means a move away from theories that necessitate a subjective understanding of what it means to possess consciousness. “We usually grant consciousness to others — of the same kind we experience in the privacy of our own mind — if they can tell us what they feel, or if they look and behave more or less like us” Tononi wrote earlier this year in Philosophical Transactions. But, he asks, what about those who can’t talk — infants, or brain injured patients? And what about animals?
The cognition of an octopus is decentralised and distributed throughout its bodyNOAA Ocean Explorer (CC BY-SA 2.0)
Traditionally, we’ve ascribed consciousness to organisms or structures behaviourally. If someone is awake, moving around and talking to us, we assume they’re conscious. And, in terms of research, much has been attributed to qualitative data, reported by subjects — if a subject can describe their experiences, they’re considered to be conscious.
Neural models, too, have often been used as a gold standard indication of consciousness. Every experience (seeing the colour red, for example, or hearing a specific musical note) has a corresponding neural activity. But Tononi believes that these correlations aren’t sufficient — or, at least, that they don’t tell the full story. Neural conditions in brain damaged patients, for example, can differ hugely from those in healthy subjects. So too can those in infants. And then there are cephalopods.
The neural activity of the octopus is of particular interest, with each tentacle possessing what is referred to as “embodied cognition”. This essentially means the cognition of an octopus is decentralised and distributed throughout its body, allowing each tentacles to integrate with other systems as well as act independently. Octopuses are also capable of sophisticated, learned behaviour, much of which we’d consider to be a mark of consciousness in humans.
And it isn’t just octopuses. Elephants recognise their own reflections and African grey parrots have demonstrated a level of consciousness similar to a human toddler. In captivity, dolphins have been observed having sex in front of mirrors, macaques feel empathy and shaken honeybees feel pessimistic.
THE ETHICS OF INTEGRATION
But our ethical approach to animals muddies attempts to understand and measure consciousness.
“Any scientific theory can be misused; this includes IIT in the context of decisions about what to do with patients with a relatively low ‘consciousness score’,” explains professor Chrisley. “But such misuse is not a part of the theory itself; it is completely consistent with the theory that society could decide that current consciousness score is not the sole factor to consider when making care decisions”.
“The gap between consciousness and its physical base is much bigger than between liquidity or information processing and their physical base. The gap is so big that it’s going to require a conceptual leap, not ordinary science, to bridge it.”
Neil Levy, Oxford Centre for Neuroethics
Quantifying consciousness could also have a profound impact on how we interact with, and treat, artificial intelligence. “Brain emulations — large simulations of the nervous system of animals — might matter morally just as much as the original animal,” argues Anders Sandberg, research fellow at the Oxford Centre for Neuroethics. “We do not know for certain this is true, but we better play it safe when we reach that level and try to treat our simulations with the same respect as the animals until we know for certain.”
THE LIMITATIONS OF PHI
IIT isn’t without its critics. Neil Levy, deputy director at the Oxford Centre for Neuroethics, argues the theory is grappling with unexplained links between thought and the physical world.
“Tononi is attempting to solve the hard problem — how can a physical system instantiate states that feel like something,” he tells WIRED. “This is a hard problem — say people who believe in it — because normal science won’t solve it. They contrast it with other claims about the relations between physical properties and other properties — how information processing in human brains is instantiated by physical systems.”
Artificial intelligence — even systems such as Apple’s Siri — could one day be argued as having a level of consciousnessMandy Cheng/AFP/GettyImages
“The gap between consciousness — tasting something, hearing something, feeling something — and its physical base is much bigger than between liquidity or information processing and their physical base,” Levy continues. “The gap is so big that it’s going to require a conceptual leap, not ordinary science, to bridge it”.
“Certain matrix computations would be super-conscious according to IIT. So either we have to bite the bullet and accept that operations I can do in a jiffy on my laptop are more conscious than I could ever be, or the IIT needs some kind of updating”.
“Although information integration seems to be very relevant to a lot of what consciousness is doing, the proponents have not made it clear to me why high integrated information is essential to what consciousness is,” Chrisley says. “I’m not convinced”.
IIT is far from perfect. But it does provide us with a new way of thinking about consciousness and offer a new set of questions about its mysteries. And these questions could help us bridge that conceptual leap.