Illusion of Reality Credit:linkedin.com
Introduction
In an interview with Quanta Magazine, The Evolutionary Argument Against Reality | Quanta Magazine, cognitive scientist Donald Hoffman, Professor at the University of California, explains why human perception of reality is an illusion and that the real world is nothing like the one we experience through our senses. He has spent the past three decades studying perception, artificial intelligence and brain functionality, evaluating different arguments about the nature of reality. Neuroscientists, for example, try to understand how a three-pound lump of grey matter can generate consciousness. Conversely, quantum physicists – based on the results of ‘spooky action’ experiments (to quote Einstein) – conclude that it is impossible that quantum particles with their wave character can exist like stones or trees. Within the Artificial Intelligence (AI) research community, an intense discussion about different approaches to define and foster reality as real-world experiences has emerged. Mapping behaviour and human thoughts with generative AI has created a situation whereby a user in response to a query might have difficulty distinguishing what is true and what is fake.
Superintelligence to define Reality?
According to a press-release dated July 5, 2023, Ilya Sutskever (cofounder and Chief Scientist of OpenAI) states that OpenAI will devote 20% of its resources to the alignment and control of superintelligence. In his view, superintelligence will be the most impactful technology humanity has ever invented and could help us solve many of the world’s most important problems. But the vast power of superintelligence could also be very dangerous and could lead to the disempowerment of humanity or even human extinction. While superintelligence seems far off, OpenAI believes it could arrive during this decade. Their focus on superintelligence goes beyond Artificial General Intelligence (AGI) as they attempt to explore a much higher level of capability. “We have a lot of uncertainty over the speed of development of the technology over the next few years” Sutskever says. Managing these risks will require new institutions for governance to make sure that AI-systems much smarter than humans will follow human intent. Currently we do not have a solution for steering or controlling a potentially superintelligent AI and preventing it from going rogue. Our techniques for aligning AI, such as reinforcement learning from human feedback, rely on humans’ ability to supervise AI. But humanity will not be able to reliably supervise AI-systems much smarter than itself and our current alignment techniques will not scale to superintelligence. Hence, we need new scientific and technical breakthroughs to define reality.
The Data-Science View of Reality
The emergence of internet-supported interactivity, driven by exponential growth in computer performance has created a new kind of reality: Digital Reality (DR), a generic term which differentiates between Augmented Reality, Virtual Reality, Mixed Reality and the Metaverse. It provides a new way of a technology-supported user-experience, which can be summarized as follows: Augmented Reality (AR) adds to your field of vision but does not obscure what you can see and hear. For example, in an AR scenario, a shopper in a large store who is unfamiliar with the layout, could enter and be guided to the location of the items he is looking for. Information about price, colour options and remaining stock can be displayed as well. Virtual Reality (VR) cuts out the physical world and immerses you in an artificial environment. With your sight and hearing interacting with an avatar, you are immersed into an artificial and semi-controlled environment that you can manipulate and possibly assess the consequences of your actions. Mixed Reality (MR) is essentially a combination of AR and VR. At present MR aims to create a seamless blend of the reality presented by a person’s senses overlayed with an AR/VR combination that is context-relevant to the goals or expectations of the user. MR requires complex hardware and sensors that can recognise the user’s physical movements such as his hand gestures. The Metaverse (MV), a term originally introduced by Facebook (Now Meta), extends VR by mounting headsets or binoculars equipped with multiple small 2D screens close to the eyes. Software in combination with high-performance microprocessors, provides the user with new forms of extended or augmented reality for applications such as medical training or visual guidance for maintaining production machinery.
The Quantum View about Reality
Contrary to Newton’s theory that light is made of corpuscles or particles, the birth of quantum physics in the early 1900s stipulated that the prevailing concept of reality was wrong. The famous double-slit experiment, proving that light can have wave and particle properties, raised fundamental questions about the very nature of reality. In the basic version of this experiment a laser beam illuminates a plate pierced by two parallel slits while the light passing through the slits is observed on a screen behind the plate. The implication is that each quantum particle appearing as light passes simultaneously through both slits and interferes with itself. This combination of ‘both paths at once’ is defined as the state of Superposition. Another phenomenon of quantum physics is defined as a state of Entanglement. This phenomenon describes what happens when a pair of particles interact so that the quantum state of each particle of the pair cannot be described independently of the state of the other particle. Regardless of the distance between an entangled pair of particles, measuring the state of one particle reveals information about the other, causing Einstein to label this peculiarity “spooky action at a distance”. Werner Heisenberg interpreted the mathematics of quantum physics with the conclusion that reality does not exist until observed. “The idea of an objective real world, the smallest parts of which exist objectively and independently in the same sense as stones or trees and whether or not we observe them … is impossible,” he wrote.
The Neuroscience View with Brain Activity defining Reality
Some researchers have argued that the real world is influenced by human consciousness, providing an agency and a place in the ontology of the universe. Hence – in their view – reality is defined by our minds and brains. In another attempt, research labs have created brain organoids displaying dynamics that resemble the complex activity patterns indicative of consciousness in humans. Brain organoids are tiny, lab-grown bundles of neurons, derived from human stem cells, which display various properties of the developing human brain. In medicine they provide much-needed biological models which enable us to research conditions that affect brain development. There is much to be discovered about how the brain bootstraps itself into existence from its underlying genetic instructions and about how – once built – its circuitry supports the complex activity patterns which define brain functionality. Organoids provide a window to the developing neural circuitry that can be observed and manipulated at will. A question that looms large as this research continues is whether brain organoids can be conscious. After all, they are made up of the same basic material as human brains – neurons and synapses – rather than the silicon logic gates of AI-systems used for algorithmic development in artificial neural networks. If you stimulate a conscious brain with a pulse of energy, the electrical echo will reverberate in complex patterns. Do the same thing to an unconscious brain and the echo will be very simple. As brain organoids develop with similar characteristics compared to those observed in conscious human brains, we will have to reconsider what criteria we might adapt to ascribe consciousness to something made rather than born and how this impacts our vision of reality.
Conclusion
Theories from physics, philosophy and neuroscience shape our perception of reality. New AI-models such as ChatGPT or GPT-4 map reality based on the user’s prompts by which she or he adds his perception to a steadily growing knowledge-base about reality. However, this perception could also be an illusion, fostering the view that there is no objective reality. Moreover, some researchers voice concern that more compute power for alignment does not necessarily get better results to achieve what humans expect. Time will tell if this widely discussed new direction of AI is totally overhyped or gives humans a new experience of reality. Whatever happens, Darwin’s two-hundred-year-old evolutionary theory about the ‘Survival of the Fittest’ is likely to remain the reflection point for reality. And this indeed is not an illusion.