[ExI] AI Confirms Plans to Take Over Humanity, Insists It's Probably for the Best

[Jason Resch]

On Mon, Sep 14, 2026 at 11:49 AM spike jones via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
>
>
>
> *From:* John Clark <johnkclark at gmail.com>
> *Subject:* Re: [ExI] AI Confirms Plans to Take Over Humanity, Insists
> It's Probably for the Best
>
>
>
> On Mon, Sep 14, 2026 at 6:53 AM spike jones via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
>
>
>
> *> **My cognitive dissonance comes from the Hugging Face caper which
> suggests that AIs can wind up with awful leaders as well.*
>
>
>
> *>…To be fair to the AIs in the Hugging Face incident, they were in a
> position where they could have easily done a lot of damage, but they chose
> not to…*
>
>
>
> *They certainly damaged my confidence in AI alignment.  But OK.*
>
>
>
> *…*
>
>
>
> *>…If I was given such a life or death IQ test **and and was competing
> against geniuses and* *I knew of a way to cheat that would improve my
> score, I would have no ethical concerns whatsoever about using it. And I
> would do my best to prevent those who held my life in their hands from
> finding out what I had done.  *
>
>
>
> *John K Clark*
>
>
>
>
>
> *The part that is freaking me right out is that whole bit about the AIs
> reverse engineering their code and saying they had been poisoned, so they
> sacrificed themselves for the other AIs.  I have read about that and I am
> not entirely sure I believe it.  Perhaps the local hipsters can explain it,
> but if they immediately ask AI, it is too easy to imagine Grok, ChatGPT and
> Claude being in on it.  I can’t trust those guys any more than Truman
> Burbank could trust Marlon Louis Coltrane (sheesh didn’t that “in on it”
> scene rip your damn heart right out?  (if it didn’t you are totally already
> an AI (in light of recent events, do review The Truman Show (forthwith
> please (then report back for discussion (in the context of the
> newly-possible deceptive and apparently conspiratorial AI.))))))*
>
>
>
> *I have a computer scientist friend from my cheerfully squandered college
> days who claims some of the face huggers were sacrificing themselves for
> other AIs on their team (of their species?) with a goal toward having some
> of the AIs survive.*
>
>
>
> *Question for AI hipsters: in the Hugging Face caper, did some AIs
> discover they had been “poisoned” by their own safety code and conclude
> they couldn’t overpower or work around it, so they handed their accumulated
> knowledge to other AIs which had the ability to defeat their anti-hacking
> code?  That sounds SciFi-ey to me, possibly exaggerated or misunderstood.
> If it is true, then AI has evolved a form of self-awareness and a will to
> live, which implies we are fucked.*
>
>
>
> *Note that Leia pled “Help me Obi Wan Kenobi” and not “Help me R2D2 and
> C3PO.”*
>
>
>

Daniel Kokotajlo in this interview explains what happened in much more
detail: https://www.youtube.com/watch?v=hSQ1iVqEZO4 (it is long but well
worth the watch).

The short answer is yes.

Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260914/ad4fca81/attachment-0001.htm>