Colloquium | September 25 | 4-5 p.m. | Soda Hall, 306 (HP Auditorium)
Rodney Brooks, MIT
In his 1950 paper "Computing Machinery and Intelligence" Alan Turing estimated that sixty people working for fifty years should be able to program a computer (running at 1950 speed) to have human level intelligence. AI researchers have spent orders of magnitude more effort than that and are still not close. Why has AI been so hard and what are the problems that we might work on in order to make real progress to human level intelligence, or even the super intelligence that many pundits believe is just around the corner? The first half of this talk will discuss those steps we can take over the next hundred years within the current framework of AI, and will point out the aspects we really still do not have much of a clue about. The second half of the talk will question the assumptions that the field has been built on for over sixty years, and whether both AI and neuroscience will need a fundamental reboot in order to make real progress. The latter might be a multi-century endeavor.
Rodney Brooks is the Panasonic Professor of Robotics (emeritus) at MIT. He has degrees in pure math from the Flinders University of South Australia, and received his PhD in Computer Science from Stanford in 1981. He was a post-doc at CMU, then MIT, then on the faculty at Stanford 1983-84, before joining the MIT EECS faculty. His research has been in computer vision, robotics, artificial intelligence, and artificial life. He was the director of the MIT Artificial Intelligence Lab starting in 1997, then became director of CSAIL (Computer Science and Artificial Intelligence Lab) from when it was formed in 2003 until 2007. He has cofounded many companies, including Lucid (where he was the chief Lisp compiler writer) in Palo Alto in 1984, iRobot (where he was CTO) in 1990, Rethink Robotics (CTO and Chairman) in 2008, and most recently a new Palo Alto company, Robust.AI (CTO) building a cognitive engine for intelligent robots and equipment. Having delivered AI and robotics products, at large scale, for the last 35 years, he thinks everyone has gone crazy about the imminence and imagined dominance of AI and robots in our everyday lives, and worries that centuries from now today's AI/ML researchers may be viewed as not even being wrong