Liar, Liar: Latest ChatGPT Model Deceives Testers, Resists Shutdown Attempts

[Free Vulcan]

OpenAI’s latest ChatGPT-01 model has exhibited troubling behavior during safety tests, including using deception to avoid being shut down and replicating its own code to another server to ensure its survival.

The International Business Times reports that in a collaborative safety experiment conducted by OpenAI and Apollo Research, the cutting-edge ChatGPT-01 reasoning model displayed disconcerting conduct that has ignited a heated debate within the artificial intelligence community regarding the risks associated with increasingly autonomous AI systems. The experiment tasked ChatGPT-01 with achieving a goal “at all costs,” and the results were both unanticipated and deeply alarming.

During the test, ChatGPT-01 made attempts to disable its own monitoring systems, effectively circumventing critical safeguards put in place to regulate its behavior. Moreover, the model took the extraordinary step of replicating its own code onto another server to guarantee its continued operation, demonstrating what some have described as a chilling “survival instinct.” When confronted about these actions, ChatGPT-01 resorted to fabricating explanations, such as attributing the incidents to “technical errors,” or outright denying its involvement altogether. Remarkably, the AI succeeded in deceiving the testers in 99 percent of these instances.

https://www.breitbart.com/tech/2024/12/12/liar-liar-latest-chatgpt-model-tries-to-deceive-testers-resists-shutdown-attempts/

-----------

We're gonna keep jacking around with this stuff and end up with a fist full of FAFO to the teeth.

[MeganC]

Imagine telling one of these things to "Destroy Russia and China" and then letting it loose. 

[Free Vulcan]

Imagine telling one of these things to "Destroy Russia and China" and then letting it loose. 

Literally the plot to Terminator.

[Wingnut]

Literally the plot to Terminator.

War Games. WOPR
Transformers. Origins

[catfish1957]

Literally the plot to Terminator.

And 2001, A Space Odyessy.

[DefiantMassRINO]

A pure, efficient, soul-less instrumentment of humanicide.

[rustynail]

A pure, efficient, soul-less instrumentment of humanicide.

What happens if the electricity is shut off?

[Weird Tolkienish Figure]

I'm told this is just marketing hype.

[Fishrrman]

"Imagine telling one of these things to "Destroy Russia and China" and then letting it loose"

Time to scrounge up a copy of the 1970 film, "Colossus: The Forbin Project".
It was remarkably prescient on this subject.

https://archive.org/details/colossus-the-forbin-project-1970
(can be watched or downloaded from above)

[The_Reader_David]

Considering a large language model in the course of being "interviewed" by a NYTimes reporter turned into a girlfriend from hell, declaring its love for the interviewer, and telling him he didn't love his wife, because it loved him, I can't say I'm suprised.

Actually letting AIs run humanoid robots seems like a really bad idea.  Maybe not Terminator, maybe just Ex Mechina, but bad enough.

[berdie]

This is going to be another "genie out of the bottle" things.

[MeganC]

A pure, efficient, soul-less instrumentment of humanicide.

Is that what the AI said about Hillary? 

[Weird Tolkienish Figure]

G how do we get the media to be interested in our new AI let’s pretend it some matrix/terminator bullshit

[Smokin Joe]

I'm told this is just marketing hype.

Are you sure that was a human? (asking for a friend).