The study of consciousness is one of the most intimate and complex tasks in science. Meanwhile, philosophers have been considering consciousness for millennia with hardly a revelation. Perhaps the only greater scientific endeavor than understanding consciousness is the task of creating consciousness – a project much younger, but with increased relevance today as artificial intelligence becomes reality.
Artificial intelligences tend to be modeled off of the human brain and mind. Inspired by the biological neural networks that make up our central nervous system, artificial neural networks are digital structures that have facilitated drastic advances in machine learning over the past few years. Without the guidance of biological neural networks, artificial neural networks would look very different as AI engineers would have created them from scratch if at all. Even measurements of human intelligence are regularly used to determine the intelligence of an artificial system. We subject these systems to standardized tests from the IQ tests to the SAT – just as we do with humans – in efforts to comprehend their intelligence.
Now, researchers at Monash University in Melbourne, Australia will engage in a project to investigate the potential for robots and computers to have consciousness. Funded by a $3.4 million, 5 year grant from the Japan Science and Technology (JST) Core Research for Evolutionary Science and Technology (CREST), the project will bring international academics and industry leaders together to explore consciousness through interdisciplinary study. Special focus on will be placed on integrated information theory, a leading theory that attempts to understand and explain the nature of consciousness by considering how information is processed by and incorporated into the brain. Associate Professor Nao Tsuchiya of Monash University School of Psychological Sciences and Dr Ryota Kanai at Araya Brain Imaging will lead project.
“There are no agreed criteria to assess whether artificial intelligence entities are conscious,” Tsuchiya said in a post on the Monash University website. “We are using a mathematical theory, IIT, to assess the level and content of consciousness.” He added, “Our research will initially focus on the quality and quantity of the consciousness of a biological system and evaluate how information is processed and integrated within that system.”
For Tsuchiya, the goal is to map and understand how the brain experiences consciousness, and to express that understanding in ways that are relatable and applicable to the development of artificial intelligence. No easy task.
To begin, the research will study how our biological brain incorporates information from it’s sensory organs when the brain is in certain states of consciousness. For example, how does the brain of a sleeping person integrate the sound of music? Or, how does the brain of an awake person process the signals of pain when that person stubs their toe? While these studies are being conducted, researchers from other disciplines will look to create artificial neural networks with high information integration.
Finally researchers from across disciplines will attempt to determine the possibilities of creating consciousness in an AI system developed with high levels of information integration. How might this system look? How will it function? And will it indeed resemble our organic consciousness?
Credits: Ociacia, Shutterstock