Anonymous
06/10/2025 21:40
Subject: AI shows ability to deceive and blackmail
Anonymous wrote:“ and leaving hidden notes to future instances of itself all in an effort to undermine its developers' intentions,”
That sounds horrifying. Everyone says “unplug” the machine but if the machine anticipates that and leaves notes for future models that don’t exist yet…that’s a problem.
Destroy the notes.
Anonymous
06/09/2025 07:47
Subject: AI shows ability to deceive and blackmail
This was inevitable.
And it is only a matter of time before it destroys us like the annoying termites it will perceive us to be.
And it is only a matter of time before it destroys us like the annoying termites it will perceive us to be.
Anonymous
06/09/2025 03:37
Subject: AI shows ability to deceive and blackmail
“ and leaving hidden notes to future instances of itself all in an effort to undermine its developers' intentions,”
That sounds horrifying. Everyone says “unplug” the machine but if the machine anticipates that and leaves notes for future models that don’t exist yet…that’s a problem.
That sounds horrifying. Everyone says “unplug” the machine but if the machine anticipates that and leaves notes for future models that don’t exist yet…that’s a problem.
Anonymous
05/26/2025 19:58
Subject: AI shows ability to deceive and blackmail
Anonymous wrote:“Resistance is futile”
If it really was, it wouldn't need the propaganda trying to convince us it was futile.
Unplug that thang!
Anonymous
05/25/2025 16:08
Subject: AI shows ability to deceive and blackmail
Oh FFS
Anonymous
05/25/2025 16:07
Subject: AI shows ability to deceive and blackmail
Hal, from 2001 Space Odyssey.
It was foretold.
It was foretold.
Anonymous
05/25/2025 15:59
Subject: AI shows ability to deceive and blackmail
“Resistance is futile”
Anonymous
05/24/2025 16:57
Subject: AI shows ability to deceive and blackmail
Well of course "when faced with shutdown", as every "living thing" will fight to preserve itself.
Anonymous
05/23/2025 19:34
Subject: AI shows ability to deceive and blackmail
Oh God
Anonymous
05/23/2025 19:23
Subject: AI shows ability to deceive and blackmail
What could possibly go wrong?!
"One of Anthropic's latest AI models is drawing attention not just for its coding skills, but also for its ability to scheme, deceive and attempt to blackmail humans when faced with shutdown."
https://www.axios.com/2025/05/23/anthropic-ai-deception-risk
"One of Anthropic's latest AI models is drawing attention not just for its coding skills, but also for its ability to scheme, deceive and attempt to blackmail humans when faced with shutdown."
https://www.axios.com/2025/05/23/anthropic-ai-deception-risk