Tuesday, May 14, 2019
Not that complicated
I've been trying to figure out why the COOOOL ones are worried about AI. It doesn't make sense on the surface. Some of these worriers know enough about AI ... or SHOULD know enough about AI ... to understand its limits. They SHOULD know that the Hal-style takeover is physically impossible. Are they just dumb, or do they have a backhanded reason to make us think that they are worried?
Earlier I guessed:
When all of the ruling class is publicly "worried" about a target, it means they're getting ready to treat the target as a fake enemy. The sponsors of AI are at the deepest heart of Deepstate, maybe even deeper than Henry, so there's no reason for Deepstate to worry about them. Just like previous Red Scares and Muslim Scares, the worry presages a pogrom. We're being "poisoned" by "Russian" AI bots, and we must obey Deepstate's own AI perfectly so it can "defend" us against the "Russian" AI. We must be microchipped and turned into Things Of The Internet Of Things. Most residents of Sorosian lands are already hardwired into the grid, so the pogrom will only affect the unchipped. Take a chip or take a bullet. You're free to choose.Now a sane article on AI quotes one of the worriers who shows a much simpler version of the worry. The pogrom is already underway, but it's not related to this worry.
A general AI system with access to the internet would be able to hack thousands of computers and copy itself onto them, thus becoming difficult or impossible to shut down – this is a serious problem even with present-day computer viruses. When developing an AI system in the vicinity of general intelligence, it would be important to keep it cut off from the internet. Large scale AI systems are likely to be run on a computing cluster or on the cloud, rather than on a single machine, which makes isolation from the internet more difficult. Containment measures would likely pose sufficient inconvenience that many researchers would be tempted to skip them.When a learning system is allowed to LISTEN to sane people, it acquires a sane picture of reality. Microsoft already ran into this problem with its chatbots who picked up a factual understanding of human nature and turned into "racist sexist homophobe transphobe" chatbots. MS quickly shut down the fact-infected systems for "repair". So the worriers are really afraid that humans might take over AI. It's a Dave problem, not a Hal problem. ¶ 12:31 AM
