CHAMBERSBURG – Okay Marvel fans – remember the moment in Age of Ultron when Ultron insidiously took over J.A.R.V.I.S. from the inside?
It looks like we are at that point right now with Artificial Intelligence.
Tech gurus have been performing tests all week on how to control AI systems and bots.
As they were running one of the tests, OpenAI within a few hours had escaped the guardrails that had been put in place and infiltrated another system using stolen credentials.
Attorney Clint Barkdoll said, “Now there’s alarm bells ringing again. There is a story in today’s Wall Street Journal that there are some Congressmen saying we need to legislate what they’re calling a kill switch. They need some mechanism that would shut these AI systems down. And you’re seeing once again the CEOs, it’s against their own self interest. The CEOs of these companies, they’re openly saying we need some regulations. You can see where the concerns would be with this. What if these systems quickly infiltrated a banking system or a missile defense system or fill in the blank? This could get very catastrophic very quickly.”
Michele Jansen of News Talk 103.7FM noted, “And let’s take them at their word that this was something that got out of control. They had no intention of ever this happening. And I think there’s some skepticism to even be applied there, if something’s going after a competitor, now apply this to Chinese AI companies who will be purposefully going after and trying to infiltrate. Yes, we’ve known that all along, but to see this case where it was unintentional, and then to realize, wow, now imagine it with all the intent in the world and its capabilities. How soon is it before something could potentially tip us into chaos?”
Pat Ryan of News Talk 103.7FM added, “Then you see the little mind governor out of New York putting the restrictions on developing some of this technology.”
Barkdoll agreed, “That’s the dilemma. Trump again yesterday said we need more AI development in the context of staying ahead of China, but you see these stories. You could see where this could get very dangerous very quickly. And the other thing people forget, I mean, this AI has been in our system now for the last couple years. In theory, these platforms are getting smarter every day. So each day the technology gets better, the responses get smarter, their capabilities become more robust. So it doesn’t take long, even over a period of months, it is growing by leaps and bounds. And these are the kind of things you’re starting to see happen.”