Bob, thx for taking a reasoned look at the “AI could wipe out humanity” headline. Rogue AI might not kill humanity but it could certainly wreak enormous havoc and chaos in our increasingly digitally dependent world.
Also you briefly mention nuclear war—what about the possibility that AI gains control of a nuclear arsenal, resulting in an enormous loss of life and huge collateral damage to infrastructure?
That is a big one. Nuclear arsenals are concentrated in the hands of a small number of governments and protected by some of the most intensive security systems humans have built. That doesn’t make the risk zero, but it does make this a very different problem from an AI exploiting thousands of poorly defended systems across the internet.
My bigger concern may be AI enabling a human attacker rather than a rogue AI independently gaining control. Either way, this is exactly the kind of catastrophic risk where the severity of the outcome justifies extraordinary defenses
Bob, thx for taking a reasoned look at the “AI could wipe out humanity” headline. Rogue AI might not kill humanity but it could certainly wreak enormous havoc and chaos in our increasingly digitally dependent world.
Also you briefly mention nuclear war—what about the possibility that AI gains control of a nuclear arsenal, resulting in an enormous loss of life and huge collateral damage to infrastructure?
That is a big one. Nuclear arsenals are concentrated in the hands of a small number of governments and protected by some of the most intensive security systems humans have built. That doesn’t make the risk zero, but it does make this a very different problem from an AI exploiting thousands of poorly defended systems across the internet.
My bigger concern may be AI enabling a human attacker rather than a rogue AI independently gaining control. Either way, this is exactly the kind of catastrophic risk where the severity of the outcome justifies extraordinary defenses