What Today's Leading Philosophers Have To Say About Conscious AI

Forbes

3 days ago

TOPSHOT - A robot using artificial intelligence is displayed at a stand during the International ... More Telecommunication Union (ITU) AI for Good Global Summit in Geneva, on May 30, 2024. Humanity is in a race against time to harness the colossal emerging power of artificial intelligence for the good of all, while averting dire risks, a top UN official said. (Photo by Fabrice COFFRINI / AFP) (Photo by FABRICE COFFRINI/AFP via Getty Images)
It is an illusion. Consciousness, that is.
Earlier this month, leading experts in the realm of the mind gathered in Heraklion, Crete, for the International Center for Consciousness Studies (ICCS) annual conference. The topic for this year was Artificial Intelligence and Sentience. All the participating philosophers, psychologists, neurosurgeons, cognitive scientists and entrepreneurs agreed; the matter at hand is no longer just a subgenre of science fiction.
According to David Hulme, CEO of Conscium, a machine consciousness research group and consultancy based in London, we have approximately five years before AI becomes a fully autonomous conscious agent. Given the numerous ethical questions associated with machine consciousness, Hulme and Conscium are as invested in the guiding principles of the research as the research itself. Their principles are readily available as an open letter on their website.
https://conscium.com/open-letter-guiding-research-into-machine-consciousness/
The general public seems to agree about the imminence of conscious AI. Clara Colombatto, a University of Waterloo psychologist specializing in the perception of other minds, presented a series of empirical studies conducted to gauge the public perception of AI consciousness. In these tests, ChatGPT-4 is the constant used as AI. The results suggest that Hulme may actually be too generous in his timetable; 57-67% of those surveyed believe that ChatGPT is already conscious to some degree.
Ok, so what? Does this mean that we must brace ourselves for a Matrix-like world? The computer scientist Roman Yampolskiy, author of Artificial Superintelligence: A Futuristic Approach, thinks so. Unlike today's cybersecurity, for example, where you can manually rectify a problem (e.g., banks issue a new credit card in cases of fraud), there will be no way out when a problem arises with conscious AI. Once it becomes conscious, it is already uncontrollable. The creation and implementation of conscious AI would then have to be perfect, which according to Yampolskiy is impossible.
However, Yampolskiy would be the first to admit he is further down the doom-scale than most. Others are more hopeful; the philosopher and entrepreneur Dmitry Volkov, co-founder of the ICCS and CEO of Social Discovery Group (SDG), a global investment firm, believes the 2013 Spike Jonze film Her is a more apt prediction. This is why ten years later Volkov founded the girlfriend app Eva AI, an 'ideal AI partner who listens, and is always in touch with you.'
The idea for Eva AI came out of data analysis from SDG, one of the first players in the online dating market. Many of the users were not always willing to meet but still desire communication. The Eva AI mission statement, according to Volkov: 'We are trying to solve the problem of loneliness from the other end, not from the Tinder attraction side.'
This is where AI may be surprisingly effective. Volkov points out, 'when you disclose emotionally invaluable information with some other person, that has a real positive impact. But if you disclose it to humans or AI, it actually does not matter.
"And we found early on [in our research] that emotional disclosure to AI has an even bigger positive impact. [The users] are discussing with a degree of honesty that is higher than if they were discussing with their brother or mother, because sometimes your brother or mother is your biggest judge.'
The ethical implications of Eva AI are vast, but Volkov thinks it can be a tool for tenderness. 'If we develop an AI that is both capable of loving and capable of being loved, then it is sort of a solution. We don't kill or torture the ones we love, even if they are inferior to us…So the project that we are engaged in is actually trying to build an AI that is not the smartest thing in the world but has the emotional intelligence that's high enough to understand your feelings.'
Much of the conversation on conscious AI centers around ethics. The cognitive philosopher Andy Clark, who was honored at the ICCS conference for his contributions to the study of consciousness, calls the ethical implications of AI 'the burning question.'
Volkov has a theory why. 'Ethics,' he says, 'comes very naturally. We are social beings, and for us, it is very important to be able to differentiate between the good and the bad.' On the other hand, he believes that 'philosophers actually created the problem of consciousness. Most people get it when they put enough effort, but it's not coming naturally.'
Clark and the rest of the ICCS mainly focused on the metaphysical problems of defining and fine-tuning the concept of consciousness during their time in Crete.
In a 1994 lecture, the Australian philosopher Dave Chalmers first articulated the hard problem of consciousness. This became the seminal 1995 paper 'Facing Up To The Hard Problem of Consciousness.' In very general terms, the hard problem questions the explanatory gap between the physical wiring and firing in the brain and the subjective experience of what it is like to be you. How can something purely physical (the body/brain) give way to something nonphysical (subjective experience)?
In order to fully understand consciousness and what it may mean for AI, the hard problem must be solved. Importantly, it can be solved. There is a concept of consciousness, it is intrinsic to the subject, and it cannot be explained in purely physical terms. This is the realist view of consciousness.
As it turned out, Chalmers was all but alone in this view at the ICCS conference. Most belonged to the illusionist camp. Chalmers even quipped that the Center ought to be called the Illusionist Center for Consciousness Studies.
Illusionism as a philosophy of mind is spearheaded by the English philosopher Keith Frankish. Frankish and the illusionists believe the hard problem is asking a bad question. Consciousness can be explained in purely physical, functional, and causal terms. It is simply a series of complex predictive, reactionary, and informational processes or models.
There are some thinkers, such as English neuropsychologist Nick Humphrey, that have issues with the name 'illusionism' itself. But a key aspect of the theory is that the subjective experience central to realism still seems to exist. In other words, human consciousness is an illusion.
Looking past these semantic disagreements, illusionism appears to be the leading view in the current philosophy of mind landscape. And yet the illusionists do not always see eye to eye. Unlike Volkov, the philosopher Katarina Marcincinova of the Kempelen Institute of Intelligent Technologies in Slovakia was fearful of the possibility of AI (functional) consciousness, opining it will be 'highly dangerous and ethically problematic.' For example, she worries about counterfeit people pretending to be human. Still, Marcincinova believes the illusion of our consciousness is essential; it allows us to create a sense of purpose for life, the world, and ourselves.
Whether it is ethics or metaphysics, it is clear that the rise of artificial intelligence has pushed the work of philosophers to the forefront. According to Andy Clark, this is where philosophy ought to be. 'The generality of philosophy is key. It gives you the tools to think a little bit more clearly about any problem, and we need to think about what AI is offering.'