Inspired by his interviews with inventor Ray Kurzweil, roboticist Rodney Brooks, and sci-fi legend Arthur C. Clarke, author James Barrat was impressed by their vision of intelligent machines. He talked about how corporations and government agencies are pouring billions into developing artificial intelligence (AI), and pondered whether we can coexist with computers whose intelligence dwarfs our own. What AI has done is to put human cognition type abilities into machines, and with processors increasing in speed every year, they are growing smarter. We'll reach a milestone in the 2020s when computers are both self-aware and self-improving, and it's uncertain how that will play out, he remarked.
As we head into an artificial intelligence explosion, when "learning machines...can write their own software, and improve their own software through experience," they will eventually surpass human intelligence, he said. "Something that is marginally smarter than the average human will quickly become much, much smarter than the average human, and there's no real ceiling on that" –they could become thousands or millions of times smarter than us, he continued. This is of particular concern because it is unknown how these brilliant machines will perceive humanity-- we have the example of how humans treat animals less smarter than ourselves. "They're in cages, or zoos, or in Sea World, or they're endangered. We don't hold them in high regard...Something thousand of times smarter than we are, may not perceive that we're worth the resources we consume," he pointed out.
Barrat noted that many companies are developing AI in secrecy, to protect commercial prospects, as well as military advancements. Yet, as AI reaches human level intelligence, it's important for developers to get together, as the technology will become increasingly volatile. To avert some of the more perilous outcomes of AI, we need more scrutiny and transparency in its development, he advised.
First hour guest, reporter David Seaman reacted to the latest revelations about the NSA, such as their covert $250-million-a-year program with tech companies to insert weaknesses into the security of their products, so they can more readily break encryption codes. It's pretty unbelievable that a tax payer funded agency is spending this kind of money to make products actually less safe for customers, he commented. Further, the NSA has been intrusively collecting large amounts of data on US citizens' social connections, who they are interacting with, both online and off, Seaman detailed. We're in a transitional time now, he noted, and a movement is gathering steam against what the NSA has been doing (a non-partisan rally Stop Watching Us is planned for October 26th in Washington DC).