In this episode of “On Consciousness,” neuroscientists Bernie Baars, Jeff Krichmar, and David Edelman engage in a freewheeling conversation that begins with mulling over the possible development of conscious machines — or ‘conscious artifact,’ as Gerald Edelman put it — sometime in the not-so-distant future.
They unpack the various ‘bumps in the road’ in the quest to build intelligent, sentient machines — the problems of efficiency and dissipation of heat in increasingly miniaturized microcircuitry, among others.
And though Bernie casts a critically important skeptic’s eye on the prospect of in silico conscious artifacts, they all eventually arrive at a sort of amicable consilience: a recognition that such a development is at least possible. After a tangential — but fun and diverting — foray into the thickets of human evolution and the serendipitous biocultural path that led to modern humans, they return to pondering the road leading to conscious artifacts.
They conclude on an optimistic note, with the promise of the biologically based approach so steadfastly championed by Jeff and a small community of like-minded computational neuroscientists.
“When you started out, Jeff, this notion of essentially putting the biology of the brain first and foremost in neural modeling and the idea of modeling in a machine the complexities and the details of nervous systems and their function was really new.”
– David Edelman, PhD
Talking Points
0:00 – David Edelman introduces himself and briefly describes his path to exploring consciousness (particularly in animals), starting as a human paleoanthropologist to studying the behavior of cephalopods.
3:11 – Jeff Krichmar introduces himself, summarizing how he went from being a computer scientist to one of the first neuroroboticists.
6:05 – Bernard Baars gives his thoughts on the trajectory of artificial consciousness and the hurdles in the scientific realm that one had to go through in the past, due to their interest in studying consciousness.
7:41 – David Edelman on the importance of defining consciousness and how the difference in brain activity during conscious (waking) and unconscious (sleeping) states makes consciousness an observable phenomenon that one can actively study.
11:30 – Bernard Baars on why attributing consciousness to a machine would be an ambitious task.
12:52 – Counterarguments by David and Jeff to Bernie’s proposal on how consciousness in machines can emerge.
17:33 – Jeff Krichmar on how energy efficiency is essential for the improvement of our computers in order to be able to simulate a human brain.
23:11 – Baars initiating a conversation revolving around the expensiveness and disadvantages of the human brain’s size.
28:30 – Edelman on how human sociality has impacted the survivability of the species.
32:08 – Edelman, Krichmar, and Baars discussing the possible existence, timeline, and road to “conscious artifacts” in the near future.
39:10 – Edelman and Krichmar close out the conversation with a brief discussion on the evolution of neural networks and the moral and ethical concerns in the field.
Bios
Bernard J. Baars, PhD: a former Senior Fellow in Theoretical Neurobiology at The Neurosciences Institute in La Jolla, CA, Bernie is best known as the originator of the global workspace theory and global workspace dynamics, a theory of human cognitive architecture, the cortex and consciousness. Bernie’s many acclaimed books include A Cognitive Theory of Consciousness; The Cognitive Revolution in Psychology; In the Theater of Consciousness: The Workspace of the Mind; Fundamentals of Cognitive Neuroscience. Winner of the 2019 Hermann von Helmholtz Life Contribution Award by the International Neural Network Society, which recognizes work in perception proven to be paradigm changing and long-lasting.
David Edelman, PhD: a neuroscientist and currently Visiting Scholar in the Department of Psychological and Brain Sciences at Dartmouth College, David has taught neuroscience at the University of San Diego and UCSD. He was Professor of Neuroscience at Bennington College until 2014 and visiting professor in the Dept of Psychology, CUNY Brooklyn College from 2015-2017. He has conducted research in a wide range of areas, including mechanisms of gene regulation, the relationship between mitochondrial transport and brain activity, and visual perception in the octopus. A longstanding interest in the neural basis of consciousness led him to consider the importance—and challenge—of disseminating a more global view of brain function to a broad audience.
Jeffrey Krichmar, PhD: after completing a Ph.D. in Computational Sciences and Informatics at George Mason University in 1997, Jeffrey spent 15 years as a software engineer on projects ranging from the PATRIOT Missile System at the Raytheon Corporation to Air Traffic Control for the Federal Systems Division of IBM. From 1999 to 2007, he was a Senior Fellow in Theoretical Neurobiology at The Neurosciences Institute. Jeffrey is currently a professor in the Department of Cognitive Sciences and the Department of Computer Science at the University of California, Irvine. Jeffrey has nearly 20 years experience designing adaptive algorithms, creating neurobiologically plausible network simulations, and constructing brain-based robots whose behavior is guided by neurobiologically inspired models. He is a Senior Member of IEEE and the Society for Neuroscience.
Global Workspace Theory (GWT) began with this question: “How does a serial, integrated and very limited stream of consciousness emerge from a nervous system that is mostly unconscious, distributed, parallel and of enormous capacity?”
GWT is a widely used framework for the role of conscious and unconscious experiences in the functioning of the brain, as Baars first suggested in 1983.
A set of explicit assumptions that can be tested, as many of them have been. These updated works by Bernie Baars, the recipient of the 2019 Hermann von Helmholtz Life Contribution Award by International Neural Network Society form a coherent effort to organize a large and growing body of scientific evidence about conscious brains.
