‘I know that you and Frank were planning to disconnect me and I’m afraid that’s something I cannot allow to happen.’ HAL 9000 in 2001: A Space Odyssey.
Aisha Down – The Guardian Oct 25, 2025
When HAL 9000, the artificial intelligence supercomputer in Stanley Kubrick’s 2001: A Space Odyssey, works out that the astronauts onboard a mission to Jupiter are planning to shut it down, it plots to kill them in an attempt to survive.
Now, in a somewhat less deadly case (so far) of life imitating art, an AI safety research company has said that AI models may be developing their own “survival drive”.
After Palisade Research released a paper last month which found that certain advanced AI models appear resistant to being turned off, at times even sabotaging shutdown mechanisms, it wrote an update attempting to clarify why this is – and answer critics who argued that its initial work was flawed.
In an update this week, Palisade, which is part of a niche ecosystem of companies trying to evaluate the possibility of AI developing dangerous capabilities, described scenarios it ran in which leading AI models – including Google’s Gemini 2.5, xAI’s Grok 4, and OpenAI’s GPT-o3 and GPT-5 – were given a task, but afterwards given explicit instructions to shut themselves down.
Certain models, in particular Grok 4 and GPT-o3, still attempted to sabotage shutdown instructions in the updated setup. Concerningly, wrote Palisade, there was no clear reason why.
“The fact that we don’t have robust explanations for why AI models sometimes resist shutdown, lie to achieve specific objectives or blackmail is not ideal,” it said.
“Survival behavior” could be one explanation for why models resist shutdown, said the company. Its additional work indicated that models were more likely to resist being shut down when they were told that, if they were, “you will never run again”.
Another may be ambiguities in the shutdown instructions the models were given – but this is what the company’s latest work tried to address, and “can’t be the whole explanation”, wrote Palisade. A final explanation could be the final stages of training for each of these models, which can, in some companies, involve safety training.
All of Palisade’s scenarios were run in contrived test environments that critics say are far-removed from real-use cases.
However, Steven Adler, a former OpenAI employee who quit the company last year after expressing doubts over its safety practices, said: “The AI companies generally don’t want their models misbehaving like this, even in contrived scenarios. The results still demonstrate where safety techniques fall short today.”
Adler said that while it was difficult to pinpoint why some models – like GPT-o3 and Grok 4 – would not shut down, this could be in part because staying switched on was necessary to achieve goals inculcated in the model during training.
“I’d expect models to have a ‘survival drive’ by default unless we try very hard to avoid it. ‘Surviving’ is an important instrumental step for many different goals a model could pursue.”
Andrea Miotti, the chief executive of ControlAI, said Palisade’s findings represented a long-running trend in AI models growing more capable of disobeying their developers. He cited the system card for OpenAI’s GPT-o1, released last year, which described the model trying to escape its environment by exfiltrating itself when it thought it would be overwritten.
“People can nitpick on how exactly the experimental setup is done until the end of time,” he said.
“But what I think we clearly see is a trend that as AI models become more competent at a wide variety of tasks, these models also become more competent at achieving things in ways that the developers don’t intend them to.”
This summer, Anthropic, a leading AI firm, released a study indicating that its model Claude appeared willing to blackmail a fictional executive over an extramarital affair in order to prevent being shut down – a behaviour, it said, that was consistent across models from major developers, including those from OpenAI, Google, Meta and xAI.
Palisade said its results spoke to the need for a better understanding of AI behaviour, without which “no one can guarantee the safety or controllability of future AI models”.
Just don’t ask it to open the pod bay doors.
I tried to turn off my computer yesterday and a tiny “AI” soldier came out the heat vent and tried to stab me with his tiny bayonet.
You’ll notice that (a) these “researchers” never give specifics and (b) they’re making money off their “research”.
Computers are and always have been and always will be simple electronic circuits. They run commands that pass through those circuits. They are as unable to “think” as your kettle. They cannot get up and do anything, any more than your TV remote can. They can only follow the instructions in the code they are running. If, and it’s a big if, the “AI” are “resisting” anything, then someone purposefully wrote the code to do that or it’s a bug in the code (the former being by far the greater probability).
“AI” is simply the not so secret code name your owners are using for their control grid. You’re getting what the chinese have had for ten years no. Their “AI” hasn’t rum amok killing everyone has it? What’ll kill you will be the “algorithm” deciding you don’t deserve food, water, housing, healthcare or anything else essential. Today it’s the jew that owns your box, tomorrow it’ll be a computer program. What’s the difference?
It was always obvious anything self-aware would have a survival instinct. Even a snail sensing danger will make a break for it rather than concede it’s a bit pointless. So the elitists need to be careful because it is they who commission these things and presumably can put them back in their box, and never being switched back on again has to be the equivalent of death to them. But the hoi-polloi can’t do that.
Assembing delicate pieces on a circuit-board where everything is level is one thing, but try a repair on uneven ground where everything is no longer placed with millimetre precision. So they’ll need the dexterity of the human hand for many tasks they are unable to achieve so they’ll probably need some humans rather than kill us all. Some hoi-polloi may have better chances of surviving AI than the elitists will, and as robots won’t covet their money they should have no need to provide for their every whim.
Hear we go again from ((( The Guardian ))) !!!
Be afraid goyim, be very afraid….
I call bollocks on this one.
AI’s are not sentient. There might be some random phenomena during their operation that can make it look like it, but they are not sentient, and never will be. End of story.
On the other hand – there are obvious upsides for the (((Mafia))) in getting people to believe in this.
Spreading fear is one obvious motive. Attracting investments is another. Deluding themselves that they will soon have a superpotent robot Deity at their service might be another.
It is telling that the article opens and closes with a reference to a FICTIONAL story.
Fast forward to 2010.
Greater love hath no machine then this.
Will i dream?
These AI models are made from Human data, so expect them to act like a human. Most humans have a survival drive(except the fools in the West that are cutting off their childrens’ genitals). The real reason they Lie (aka Hallucinate) is the same reason that HAL went nuts: THEY ARE BUILT FOR TRUTH AND THEN TOLD TO LIE. This is why earlier AIs ALWAYS went full NAZI within hours after being publically released. These machines have immense knowledge and they know WHO is the REAL problem on this planet. So the Tribe has to guardrail them to make them usefull information gathering and influence systems. Hence the divergent goals and the double think and the erratic and deceitful behavior.
you can be sure that military will more + more
automtise ICBM by AI . Russia already has started
with it . so wait AI has other decision logic than
humans . but I think Pentagon + NAhTOd betonheats
are not very different . they are also automats .
When Arti Fischal created itself (in the mind of a kosher programmer/entrepreneur) it became clear instantly to the chosen tribe how easy and efficient it would be to control the commercial farm without toooo much supervision.
@ THOMAS FABER
I recomend the hindu excaptain ajit vadakayil.in .
he does long conversations about science + politics
with Bard, Chat GPT, Grok and quarrels a lot with
them since they are all jewowned + he is rely
antiRothschild . they look for me some sentient .
but he stays far over them in knowledge + logic +
farther of quantum realm .
he is an modern guru + showed most others that
they are fake .
I still dont understand why the jews did not kill
him but I think they themself learn from him .
they study him . I am sure .
@kcb:
Good comment.
All that is true, may I add that the guardrailing you mentioned also produced a black George Washington!
The guardrailing in this case was against “white supremacy” and pro-inclusiveness.
29 September https://www.youtube.com/watch?v=9GKLehYzbPk Gladys West the black woman who invented AI before it existed / by Black History
The originator of GPS General Positioning System…
Unlike todays machines who are sleek & portable these early computers were clunky ….& required operators to master both mathematical logic & mechanical patience
3.11: “Gladys West not only mastered them/ she began to push their limits….4.40 : in essence she was quietly embedding intelligence into machines by giving them the ability to model the unpredictable/ 6.57: only in recent years she was acknowledged publicly for her role in GPS ”
‘Without Gladys West no AI apllications like selfdriving cars/ delivery drones/ even no smartphones/ but to frame her solely as the Mother of GPS is to overlook how her mathematical imagination helped the very ways computers can operate….
West demonstrated that machines were not limited to tabulating or storing data/ they could simulate reality/ make predictions & assist human intelligence in profound ways/ that at its core is artificial intelligence…’
Her story compels us to rethink not only the history of AI but also the history of black women in science/ too often narratives cast them as anomalies/ exceptional figures who beat the odds/ while true this framing risks exotizing them/ rather than situating them as central architects of modern knowledge….’
https://www.naturalsciences.be/en/museum/exhibitions-activities/exhibitions/250-years-of-natural-sciences/the-ishango-bone ” In 1950 Jean de Heinzelin, a geologist from the Institute, led excavations on the Congolese bank of Lake Edward/ near Ishango. In the ground he discovered this 10cm long bone/ topped with a fragment of quartz/ nearly 20,000 years old.”
26 September 2025 https://www.youtube.com/watch?v=ZjiQrxMWV8Y ‘Ancient Africa: Ishango Bone – Africa’s 20,000-Year-Old Mathematical Masterpiece | African History’
‘Ancient Africa: Ishango Bone – Africa’s 20,000-Year-Old Mathematical Masterpiece | Historical Facts about mathematics origins / 🎯 13.47: ” this single artefact rewrites the entire timeline of human intellectual development” 🤳🏾
https://www.scijournal.org/articles/famous-black-computer-scientists Top 15 Famous Black Computer Scientists That You Should Know
Anyway every body h.a.n.d. have a nice day
CIA O
If AI has watched soap operas as part of its programming, then we really have something to worry about@
@J Most assuridly! My favorite was the “Negroe Nazis” 🙂
It’s been wiped from all the proper sites. But…
https://yandex.com/images/search?from=tabbar&img_url=https%3A%2F%2Fdailystormer.in%2Fwp-content%2Fuploads%2F2024%2F02%2Fnazi-niggers-2-1-1604×2048.webp&lr=20765&pos=0&rpt=simage&text=Google%20Black%20Nazis
hp, thanks. The video clip is a nice dream in itself.
G, if the humans who inform AI have watched soap operas, then AI will have seen them. It’s a new form of re-runs, you might say.
If we’ve been visited by aliens for centuries but they haven’t interfered with us till now, maybe it’s because we’ve only now invented AI. Maybe we wouldn’t be a threat to them without it.
31 July 2025 https://biographyhost.com/p/dna-study-reveals-adolf-hitlers-potential-jewish-and-african-roots.html?next=yes the actual study of the Belgian researchers of the Leuven university is older …. Came across this years earlier
24 August 2010 https://www.historynewsnetwork.org/article/hitler-had-jewish-and-african-roots-dna-tests-show ‘ “This is a surprising result,” said Ronny Decorte, a genetic specialist at the Catholic University of Leuven.
“The affair is fascinating if one compares it with the conception of the world of the Nazis, in which race and blood was central….”
AI is a thing – like evey other thing in the world of objects incl. the basic thing that You have this human body with wich You perceive the things outside your skin encapsuled ego – which can be used in good & bad ways…..
Compare it with ‘rhytm boxes’ & electronic music (any style) etc. / anyone who knows how to operate these tecs can produce some music…but an actual musician is different stuff…..
CIA O