How neuroscience and artificial intelligence challenge ideas of what makes humankind special
DUBLIN — Time and again in recent centuries, science has challenged humankind’s conception of itself. Copernicus demoted humans from their central location in the cosmos. Darwin denied their inherent distinction from other animals. Computers now trounce human intellect, defeating the smartest people at chess and on Jeopardy! And lately modern neuroscience has pretty much demolished traditional views of souls and selfhood.
“The neuroscience revolution and the genomics revolution are changing the way that we look at ourselves,” says neuroscientist-science writer Lone Frank of the Danish newspaper Weekendavisen. “If you take neuroscience as an example, it’s becoming very clear that we’re killing off the soul. We are realizing, and it’s seeping into our culture, that we are our brains. And that means also that there is no essential self in there.”
Frank, author of the recent books Mindfield and My Beautiful Genome, acknowledges that many people don’t especially appreciate hearing that message. But it’s a natural enough topic to discuss on occasions where science meets culture, such as the Euroscience Open Forum 2012. Known as ESOF, the conference takes place every other year in a European city eager to host leading scientists and science journalists for a mix of technical presentations and culture-oriented discussions and activities.
During a talk and at a news conference July 12, Frank remarked on neuroscience’s impact on humans’ outdated notion of self.
“There’s no true one self in any person,” she said. “We are our brain — that means we are the state that our brains are in at a certain time. And it’s also becoming very clear from neuroscience that the brain is extremely plastic and that we can change it in different ways, with devices, with cognitive techniques …, with drugs.”
In other words, you don’t have to be stuck with who you think you were born to be. “Instead of looking for who we are, we are trying to figure out who we want to be,” Frank said.
Assaults on humans’ ideas of identity are coming from another angle as well, namely rapid progress in the field of artificial intelligence. As computers outcompete humans in many intellectual pursuits, people have clung to certain sensory skills as evidence of superiority. This strategy may turn out to be of special interest to other occupants of the planet, suggested Brian Christian, author of The Most Human Human: What Artificial Intelligence Teaches Us About Being Alive.
“In some ways the development of artificial intelligence may prove to be an unexpected boon for animal rights,” he said.
Comparing ourselves with machines to define humanness is a shift from two millennia of philosophical effort to contrast humans with other animals, Christian pointed out.
“We’re finding that precisely the things that seem to most greatly differentiate us from artificial intelligence are the very types of skills that we share with other animals — the ability to live the sort of sensory-embodied existence where we recognize objects, navigate space, exchange communication … a lot of the things that philosophers had written off about the human species precisely because animals shared them. These have turned out to be the greater hurdles for artificial intelligence.”
Someday, of course, machines may close the gap between them and humans (and other animals). Humans are still better than computers at driving cars, for instance, because of the need to quickly identify objects appearing in the road — distinguishing a shadow from a piece of trash or a dog or a child. Computers may achieve such skill, Christian believes, through programming that makes use of Bayesian probability.
“The first several decades of computer science were defined by the rigidity of the logic — either it was zero or it was one, it was true or false,” Christian said. “I think the next few decades are going to be defined by probabilistic reasoning using Bayesian inference.”
With Bayesian reasoning, inferences are based on probabilities weighted by evidence from experience. It seems to be the way humans learn how to cope with the world and reason about appropriate behaviors (see “The Probabilistic Mind” SN: 10/8/11, p. 18).
Human evolutionary success no doubt has rested largely on the brain’s ability to incorporate the principles of Bayesian mathematics. Ironically, though, about a century ago human scientists largely abandoned Bayesian methods for scientific purposes. Standard statistical methods for reaching scientific conclusions instead follow a textbook prescription concocted to make calculations more feasible, not more correct.
Nowadays, that same advanced computing power that makes artificial intelligence a threat to human egos also allows difficult Bayesian computations to be performed more efficiently, so Bayesian methods are making a comeback. It’s another benefit of the brain’s ability to change itself.