Democratic Sen. Mark Warner of Virginia was speaking a few weeks in the past about Anthropic’s synthetic intelligence “Mythos mannequin.” Warner said, “This device broke into nearly all of our labeled programs, not in weeks however in hours.”
Hmmm.
A man-made intelligence system can break into nearly all the United States’ labeled programs in hours.
That sounds fairly harmful.
Am I permitted to ask a follow-up query?
How does the Mythos mannequin do once you ask it to deal with the North Korean, Iranian, Russian, or Chinese language labeled programs?
From the sound of issues, you might give some spook a few days with the Mythos mannequin and all of North Korea’s nukes can be disabled. The Center East wouldn’t have to fret about Iranian ballistic missiles, as a result of none of them would fly.
Proper?
Am I lacking one thing?
I imply, if the issue is that our laptop doesn’t communicate Persian, that would absolutely be solved.
Possibly the pc doesn’t communicate Persian now.
However barely modify the massive language mannequin, and the pc does communicate Persian … now.
Are the programs the U.S. authorities makes use of to guard its labeled paperwork so inferior to the programs utilized by different international locations that Mythos can crack our programs, however not theirs?
In that case, disgrace on us.
However simply juice the AI by some means, and set it to work.
Give it some thought: We don’t have to fret concerning the Russian nuclear arsenal, as a result of we’ve set AI to work disabling it. We don’t have to fret concerning the Chinese language stealing our mental property, as a result of all the thieving software program has been disabled. Actually, possibly we might begin stealing Chinese language IP for constructing electrical automobiles, and photo voltaic panels, and robotics, and the like.
Shoot! Possibly AI might determine find out how to make the Lincoln Memorial Reflecting Pool blue once more.
I’m only a foolish clown at a keyboard, typing away out right here within the low cost seats.
But when I’ve had this thought — if AI can crack our computer systems, can’t it crack theirs? — haven’t not less than a couple of folks within the U.S. authorities had the identical thought?
Have these folks put the thought into motion?
Possibly the US is once more a hegemon. The entire different international locations have been infiltrated; their programs received’t work; and so they don’t understand it but.
Or possibly Sen. Mark Warner has barely misstated issues.
Heck if I do know.
Ask ChatGPT.
Mark Herrmann spent 17 years as a accomplice at a number one worldwide regulation agency and later oversaw litigation, compliance and employment issues at a big worldwide firm. He’s the writer of The Curmudgeon’s Guide to Practicing Law and Drug and Device Product Liability Litigation Strategy (affiliate hyperlinks). You may attain him by e mail at inhouse@abovethelaw.com.
The submit An Entirely Innocent Question About Artificial Intelligence appeared first on Above the Law.