Im tired of sounding either crazy or unintelligible when explaining MIRI to people, so Im toying with elevator pitches that avoid both problems. What are yalls responses to this one? The Machine Intelligence Research Institute studies how to predict and prevent risk from powerful machines that make their own decisions. When computers are able to change their own programming, they can make improvements that humans wouldnt have thought of. Unfortunately, they can also make changes that prove to be harmful overall. A gain in one area may cause a loss in another. For example, if you tell a machine to build inexpensive cars and allow it to accomplish that task however it sees fit, it may design and construct the most efficient car factory in existence--or it may use the tools on hand while monopolizing the underground market in stolen car parts. MIRI wants to ensure that advancements in artificial intelligence lead to the former kind of improvement rather than the latter.
Posted on: Wed, 30 Oct 2013 23:44:59 +0000
Trending Topics
Recently Viewed Topics
© 2015