Thursday, January 15, 2015

Lack of interest by government in regulating runaway AI exposed

So far, efforts by Etopia News to determine the views of elected California officials on the issue of regulating the development and implementation of advanced artificial intelligence systems have only revealed this is a subject not yet on the radar of those charged with protecting the public from possible negative side-effects from building software that can outperform humans at most jobs and get into all manner of other mischief, even if it doesn’t evolve into Skynet or the Matrix.

From newly-elected California State Senator Bob Hertzberg, came this reply to an inquiry about what the state legislature might be considering by way of regulating powerful computer systems capable of simulating human-level intelligence, or self-evolving into systems much more intelligent than humans:

“As for high-level intelligence, frankly, that’s not an issue that’s come up during the seven days we’ve been in session. Maybe someday soon but for now we’re tryin’ like heck to get our heads around tax reform, energy, local guvment, etc. “

Aurelio Rojas, a spokesperson for California State Senator Jerry Hill, who represents a large swathe of Silicon Valley, shared by phone the general proposition that the slowness of the legislative process, being more cumbersome than the process of technological innovation, results in some problems not being addressed until after they manifest themselves acutely.  In the case of a self-aware computer system, waiting until after it’s operational might mean it’s too late to do anything about it.

Rojas also pointed out that electeds tend to respond more actively to issues that are raised by organized groups of their constituents, and that no one has yet come to Senator Hill asking that he look into the questions of regulating super-intelligent computers or preventing a runaway AI.  No one is asking for public hearings on the subject or asking for specific safeguards against powerful AI systems that could and might routinely violate personal and commercial privacy and cause other negative effects.

There is massive public concern about privacy violations via computer-based surveillance, but the opposition to this surveillance is based on worries about the privacy violations, not the existence of the AI systems that make them possible.  This is only one of the things that powerful AI can get up to.  Greater interest from electeds could go some way towards igniting the effort to regulate runaway AI, even if they haven’t yet received calls and e-mails from constituents urging them to look into this emerging threat. 

It is, in short, a tremendous opportunity to demonstrate exactly those qualities of foresight, understanding, and action that constitute true leadership. .

No comments: