Back in March, New York University hosted a symposium in New York City called Canonical Computations in Brains and Machines, where a group of neuroscientists and experts in the field of artificial intelligence spoke about overlaps in the ways in which human beings and machines think and process information. According to one of these neuroscientists – Zachary Mainen of the Champalimaud Centre for the Unknown – we might expect advanced machines to soon be able to experience some of the same mental problems that people do.
“I’m drawing on the field of computational psychiatry, which assumes we can learn about a patient who’s depressed or hallucinating from studying AI algorithms like reinforcement learning. If you reverse the arrow, why wouldn’t an AI be subject to the sort of things that go wrong with patients?” Mainen said.
The neuroscientist went on to explain that mental illnesses like depression and hallucinations depend on a chemical in the brain called serotonin, and if serotonin is being used to help intelligence systems solve a more general problem, “then machines might implement a similar function.” In other words, if serotonin can go wrong in humans, it could also go wrong inside of a machine.
To most, the idea of artificial intelligence advancing to a point where it actually suffers from many of the same emotional problems and mental illnesses as human beings do is unnerving at the very least. They are right to feel this way, frankly, because the more artificial intelligence advances, the more difficult it will be for human beings to control them. It sounds like something straight out of a science fiction film, but this is quickly becoming a very real concern that society needs to take seriously. (Related: A new AI has the ability to teach itself with a reinforcement learning algorithm that results in superhuman abilities within hours.)
According to Katja Grace of the Future of Humanity Institute at the University of Oxford, artificial intelligence will begin to exceed human performance within the next few decades as machines continue to expand their capabilities. In order to arrive at this conclusion, Grace surveyed a number of leading researchers in artificial intelligence (a total of 1,634 to be exact) and asked them when they expect artificial intelligence to start outperforming human beings. The findings indicate that this may happen much sooner than we think.
The experts that responded to the survey predict that AI will outperform humans within the next ten years in tasks that include translating languages (by the year 2024), writing high school essays (by the year 2026) and driving trucks (by the year 2027). However, other tasks, such as working in retail (2031), writing a bestselling book (2049) and working as a surgeon (2053) are expected to take significantly longer for machines with artificial intelligence to master. (Related: Here are eight professions that are expected to see human beings replaced with artificial intelligence technology in the years ahead.)
Furthermore, the experts predicted that there’s a fifty percent chance that artificial intelligence will be better than human beings at mostly everything in just four and a half decades. Relatively speaking, that’s not a very long time, and human beings had better make plans to adapt to this inevitable technology-driven world or risk becoming obsolete. See Robots.news for more coverage of robotics and AI.
Sources include:
Please contact us for more information.
