By quite literally unplugging them. Unless there is a system out there I am unaware of, they all need electricity to run.
Now what you raise here is an issue of proliferation and is a different issue entirely.
Both cause irrational fear but have entirely different scenarios. We shouldn’t muddle the two together.
As far as any kind of AI proliferation ban that might be thought of as similar to agreements on non proliferation of nuclear weapons, well that’s just not possible.
Unlike nuclear power, AI research doesn’t require any kind of unobtainable or strictly controlled substance. Governments could try and restrict it all they want but there is nothing preventing one person or a small team in a garage from making world changing breakthroughs.
I don’t find this cause for alarm, in fact I think this is great. The more R&D is distributed the more we will understand and have disparate knowledge of how to build and direct reliable AI systems. There will always be big money players that will dominate but if we’ve learned anything from the last 70 years just starting out with money and resources doesn’t guarantee a company will survive or not be supplanted by an upstart we have yet to hear anything about.
What I’m far more concerned about is people leveraging public fears, grown out of decades of AI apocalypse science fiction, to trick the public into supporting unnecessary, pointless, and unenforceable restrictions on a technology in order to attempt to build moats around first movers and entrenched players. That’s a much greater and immediate danger to the general wellbeing of society.
Most of the current fear around GAI and SGAI is not only so far ahead of the current state of the art as to border on delusion and paranoia but comes from people and institutions with very suspect motivations. ( not saying you personally, just in general )
It’s all good philosophical material and makes for great debate but these concerns are no more concrete now then they were 30 years ago. We’ve just finally got our hands on a few tools that makes us think we are in the cusp of great leaps. But right now there are far too many people trying to not just count but boil and fry their chickens before the hen has even laid the egg.
