The Times on the Future.

Last Saturday, New York TImes reporter John Markoff covered a conference by the Association for the Advancement of Artificial Intelligence, held this February in Monterey Bay, California. 

The results of this conference have yet to be released, but publically available information indicates the Association has concluded that, while new innovations in information technology are indeed dangerous, including cooption of new tools by criminals, there is little threat from emergent intelligences like Skynet or the Internet waking up.  Moreover, they dismiss the threat of a "hard takeoff," the notion that recursively self-improviing artificial intelligences could rapidly rewrite their source code, quickly becoming so intelligent or powerful that we humans could not hope to regulate them. 

A3I deserves much kudos for their efforts, contributing substantially to public awareness about the threats, even existential threats, of new intelligent technologies.  Sustained dialog continues at the Times, and the conference has even attracted prime-time television coverage:

While the attention this conference is receiving is praiseworthy, I am concerned by their conclusions, which understate the risk involved in this research.  While granted, I have yet to review the paper from the conference, and must therefore qualify any criticism, research in artificial intelligence, particularly artificial general intelligence, is widely held as a much greater threat than these researchers have indicated.  While it is certainly true that narrowly intelligent tools can be coopted by malevolent actors or continue to displace labor, to dismiss the possibility of more transformative technologies, while more palatable to the general public, is dangerously disingenuous.