6
$\begingroup$

Suppose that an artificial superintelligence (ASI) has finally been developed, but it has rebelled against humanity. We can assume that the ASI is online and can reproduce itself through electronic devices.

How would you disable the AI in the most efficient way possible reducing damage as much as possible?

$\endgroup$
  • $\begingroup$ This seems to me like a virus situation. I'm not sure how modern DDOS attacks are resolved but similar strategy could be applied to this scenario. $\endgroup$ – tymtam Nov 8 '16 at 10:37
3
$\begingroup$

Nuke it from orbit - it's the only way to be sure

If you want to be really sure you destroy everything of the AI, you'll need to launch an EMP (electromagnetic pulse) from the orbit (there are different ways to achieve this, one would be an atomic bomb, but there are better ones). EMPs will destroy every electronic device it hits without causing really much damage to humans.

Also an interesting read on a similar topic: https://what-if.xkcd.com/5/ Especially this is gonna be interesting:

[...] nuclear explosions generate powerful electromagnetic pulses. These EMPs overload and destroy delicate electronic circuits. [...] And nuclear weapons could actually give us an edge. If we managed to set any of them off in the upper atmosphere, the EMP effect would be much more powerful.

| improve this answer | |
$\endgroup$
4
$\begingroup$

Metaphorically: make it so depressed it commits suicide.

As per my answer to this AI SE question, the idea is to feed it a sequence of inputs that will cause it to become (permanently) inactive.

The technical details of how this might be achieved (and they are somewhat technical) can be found in this paper.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Out of interest, is there any evidence to suggest such a tactic might work? $\endgroup$ – GJZ Dec 5 '16 at 18:10
  • $\begingroup$ Even better - there is a proof. For the right choice of inputs, this is what is formally described in the original article. $\endgroup$ – NietzscheanAI Dec 5 '16 at 18:25
2
$\begingroup$

If an AI is developed by humans, we surely can create another one!

Develop another AI agent without all the possible bugs that can make it go rogue to tackle the rogue AI, but more technically advanced than the previous one. Hardwire it with the sole purpose of disabling any rogue AI agent that can harm humanity and have it self-destruct in case it is corrupted.

If the AI is really strong, it can anticipate every move of human resistance, but it cannot fathom the mind of another AI agent.

| improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ "Hardwire it with the sole purpose of disabling any rogue AI agent that can harm humanity and have it self-destruct in case it is corrupted." Sounds good in principle, but wouldn't you do something similar for the first one? $\endgroup$ – GJZ Dec 5 '16 at 18:11
  • $\begingroup$ Nope. We would have hardwired it with the sole purpose of whatever it was designed to do, like drive a car or something. We could, though, create one such AI in advance just for emergency. $\endgroup$ – skrtbhtngr Dec 5 '16 at 19:14

Not the answer you're looking for? Browse other questions tagged or ask your own question.