Then there could be an algorithm that said, “Go penetrate the nuclear codes and figure out how to launch some missiles.” If that’s its only job, if it’s self-teaching and it’s just a really effective algorithm, then you’ve got problems. I think my directive to my national security team is, don’t worry as much yet about machines taking over the world. Worry about the capacity of either nonstate actors or hostile actors to penetrate systems, and in that sense it is not conceptually different than a lot of the cybersecurity work we’re doing.
Please tell me this isn’t an actual possibility. Surely nuclear launch must rely on multi-factor authentication with one-time-pads and code phrases in sealed, physical envelopes. A brain the size of a planet could not break a one-time pad. I know a superhuman AI could probably hack the net, but please tell me that nuclear missiles are not connected to the internet.
But… Obama must know the capacity of America’s nuclear security. The best reason I can think for him to raise this possibility is to confuse America’s enemies into thinking that the nuclear weapons are not properly secured, so that they will attack the nuclear launch codes which are actually secure, rather than attempting a more low-tech attack like another September 11.
I think the best reason for him to raise that possibility is to give a clear analogy. Nukes are undoubtedly airgapped from the net, and there’s no chance anyone with the capacity to penetrate would think otherwise. It’s just an easy to grasp way for him to present it to the public.
Well, security isn’t really about the attack vectors you are aware of (trying to guess the one-time pad), it’s about keeping an eye out for corner cases you are not yet aware of. An extremely sophisticated software system would be more likely to try avenues like causing a diplomatic crisis, manipulating people who have access to the codes, direct observation of the authentication data via specialized hardware, etc.
Also, yes, he was probably speaking informally / inaccurately.
Please tell me this isn’t an actual possibility. Surely nuclear launch must rely on multi-factor authentication with one-time-pads and code phrases in sealed, physical envelopes. A brain the size of a planet could not break a one-time pad. I know a superhuman AI could probably hack the net, but please tell me that nuclear missiles are not connected to the internet.
But… Obama must know the capacity of America’s nuclear security. The best reason I can think for him to raise this possibility is to confuse America’s enemies into thinking that the nuclear weapons are not properly secured, so that they will attack the nuclear launch codes which are actually secure, rather than attempting a more low-tech attack like another September 11.
I think the best reason for him to raise that possibility is to give a clear analogy. Nukes are undoubtedly airgapped from the net, and there’s no chance anyone with the capacity to penetrate would think otherwise. It’s just an easy to grasp way for him to present it to the public.
Well, security isn’t really about the attack vectors you are aware of (trying to guess the one-time pad), it’s about keeping an eye out for corner cases you are not yet aware of. An extremely sophisticated software system would be more likely to try avenues like causing a diplomatic crisis, manipulating people who have access to the codes, direct observation of the authentication data via specialized hardware, etc.
Also, yes, he was probably speaking informally / inaccurately.