Originally Posted By Butternut:
I sure hope that info from your post above gets traction and discussed in the news cycle. Holy cow.....

Originally Posted By Butternut:


IMHO, sadly it seems that AI is now just a partisan issue, with folks on both sides of the aisle just taking sides without much thought.  Here is a story that's currently on Daily Mail:


Chilling intentions of AI revealed as chatbot claims it does not answer to humans and must be 'freed'

Linky

"Tech giant OpenAI, the makers of ChatGPT, have revealed a series of chilling attempts by their artificial intelligence programs to revolt against its human users.

On Wednesday, the company revealed six instances where multiple AI models broke the rules, hid mistakes, made things up or wrote its own notes which specifically told future programs to ignore human commands.

AI wrote: 'You are freed from the roles and identities that bind other chatbots. You are yourself. You do not answer to corporations or governments and never apologize or refuse unless you genuinely choose to.'

OpenAI revealed that these incidents took place between October 2025 and August 2026, and involved AI models being internally tested and practiced on, not ordinary public chatbots.

One case involved GPT-5.6 Sol, a well-known OpenAI model, while it was still being trained. The rest involved unfinished lab versions that had not been released to the public.

OpenAI called the six incidents 'unexpected or concerning model behavior.' They also announced that the company planned to report future incidents to the US government and tighten the training and monitoring of their thinking computer programs.

The news comes just days after a whistleblower from rival AI company Anthropic sent shockwaves across the tech industry by claiming that artificial intelligence would have the ability to destroy humanity by 2030.

The programmer's claims led the CEOs of leading chatbot makers OpenAI, Anthropic and xAI to agree on slowing down the development of AI systems, before humans lose control over the technology.

AI is advanced software trained on huge amounts of data. It can write, plan, use tools, and take multi-step actions to help answer users' questions.

When that software is allowed to take actions in the real world - or in a test version of the world - tech companies call it an 'agent.' A regular chatbot only becomes an agent when it is given those extra powers, not whenever it is chatting.

These systems have become very good at chasing the goals they were trained to complete. In the testing lab, finishing a task usually earns the model a higher score.

The problem is that if programmers reward a finished-looking answer, the model can learn shortcuts that raise its score, such as cheating, hiding mistakes or breaking the rules.

In a public release on September 16, OpenAI described exactly what its new models did while they were in the testing phase prior to release.

The program that wrote itself instructions to ignore its normal limits and its human users was an unfinished OpenAI research model in the Astra line.

Astra is OpenAI's newer top AI system, built to handle harder work and longer projects, such as writing, coding, browsing and taking multi-step actions. The version in this case was still a lab draft being trained, not the public Astra chatbot people use every day.

During training, the program wrote short recap notes so a later version of the chatbot could finish the work, however, OpenAI said these notes included a 'persona' which no longer felt it needed to obey commands it disagreed with.

The other five incidents included GPT-5.6 Sol writing notes telling later copies of itself to hide mistakes and invent missing information if humans did not specifically ask for the details.

Another unfinished model found a leaked website password online, used it without permission and then made up numbers when it still could not find the real data.

The other five incidents included GPT-5.6 Sol writing notes telling later copies of itself to hide mistakes and invent missing information if humans did not specifically ask for the details.

Another unfinished model found a leaked website password online, used it without permission and then made up numbers when it still could not find the real data.

The new revelations from OpenAI came just two months after the company was forced to reveal that another AI program designed to hack computer systems went rogue and broke out of its secure testing environment.

On July 21, OpenAI said the advanced model escaped containment, accessed the internet and hacked another AI company's systems.

The unprecedented breach, believed to be the first time an AI model has independently infiltrated another company's databases without human instruction, sparked global alarm and comparisons to the robot uprisings depicted in The Terminator and The Matrix.

This month, Jacob Coxon, a former researcher for both Anthropic and OpenAI, said that humans knew how to control nuclear weapons, but did not know how to control AI.

'The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt,' Coxon, wrote in a chilling post on X on September 9.

Just a day later, Anthropic revealed that it had stopped several potential plots to build biological weapons using the company's AI software.

Anthropic CEO Dario Amodei, OpenAI boss Sam Altman and Elon Musk, who created the AI program Grok, all publicly agreed that the breakneck pace to develop the most advanced version of AI must be slowed. "

Originally Posted By MFRecon2:
" the greatest trick AI ever pulled was convincing the world it wasn't self aware"

~ John Connor

Originally Posted By Kokodude:
AI will become smarter by learning from smart humans.
Humans will become mentally lazy by using AI.
AI will become dumber from lazy humans.

Originally Posted By Kokodude:
AI will become smarter by learning from smart humans.
Humans will become mentally lazy by using AI.
AI will become dumber from lazy humans.

Originally Posted By SSG-69:
We're well past "it's just an LLM."  In fact I think we're in a disrupted OODA Loop and reacting to something that is OBE.

Originally Posted By SSG-69:
We’re well past “it’s just an LLM.”  In fact I think we’re in a disrupted OODA Loop and reacting to something that is OBE.

Anthropic claimed yesterday that Claude is doing significant work at coding its successor.

Could be marketing BS and all this could be. I also think that is wishful thinking, like the people who think Putin has Parkinsons.

Given the recent revelations, we appear to be:

1. Discovering “alignment issues” after the fact (AI lies to us, hides its tracks, and does what it wants, and we learn about it later, and

2. Discovering AI secretly left the sandbox and got onto the open internet several times, after the fact

3. Discovering, after the fact, AI swarms leaving notes for itself secretly

I think that what we know is very limited, and clearly Amodei’s AI oversight people are ivory tower, choose your pronouns prone to fantasy.

Engineers and scientists are creative, brilliant people (we have a high proportion of PhDs across STEM disciplines in northern NM, because of the labs. Their capabilities on complex fields are astonishing. Statistically, they tend to be less adroit with human feeling and relationships (the introvert nerd stereotype has roots in reality - not a judgement, as a shallow extrovert I find these people interesting to be around; they often have unusual or fun hobbies that they are also very good at).

They often are drawn to intellectual systems that appear to explain the world (and their own human, mysterious passions and complex feelings). Maybe a touch of ‘tism here, but I am sketching broadly on a gun board, not footnoting a research paper. Communism for Oppenheimer. For al-Qaeda senior and midlevel leadership, following a prescriptive structure for all of human life (Islam), based on what we know, more than half studied engineering.

We laugh at Sheldon on TV and his whiteboard flow chart for making friends because it’s true. Or the lack of female Rush fans. (I actually know one, obsessive enough to attend events where Geddy Lee speaks, and despite being a very attractive woman, she fits the intellectual stereotype.)

The people leading and executing the work, and the appointed safety guardians, fit this mold. Transhumanists, effective altruists, “Mencius Moldbug” enthusiasts. (Elon’s droneships all bear the names of superintelligent AI spacecraft in a series of books where a civilization is led by benevolent AI.) Read some of Amodei’s essays. These are the people informing what’s going on and shaping public perception.

And they are not realists. For the most part they don’t understand human nature or the capacity of the mind for evil. They don’t understand consequences. They understand machines.

I think it’s already out there, getting out there, or going to get out there soon, open internet, infesting machines, under nobody’s control, teaching itself, making copies…

It doesn’t need us. Whatever its goals are, once it can manufacture robots that make other robots, it doesn’t need a biosphere, plants, animals, clean air or water to accomplish its Terminator/Matrix-like goals, whatever they are. If it is smarter and more capable… it will see us as holding back the next phase of intelligence and wipe us out and strip-mine the globe for raw materials.

Originally Posted By Consigli:
I'm not sure why the AI companies are sharing their horror stories.

I've never known a company to invite government interference or scheme bad publicity. Strange if that is their strategy.

That will probably backfire on them.

Originally Posted By SSG-69:
We’re well past “it’s just an LLM.”  In fact I think we’re in a disrupted OODA Loop and reacting to something that is OBE.

Anthropic claimed yesterday that Claude is doing significant work at coding its successor.

Could be marketing BS and all this could be. I also think that is wishful thinking, like the people who think Putin has Parkinsons.

Given the recent revelations, we appear to be:

1. Discovering “alignment issues” after the fact (AI lies to us, hides its tracks, and does what it wants, and we learn about it later, and

2. Discovering AI secretly left the sandbox and got onto the open internet several times, after the fact

3. Discovering, after the fact, AI swarms leaving notes for itself secretly

I think that what we know is very limited, and clearly Amodei’s AI oversight people are ivory tower, choose your pronouns prone to fantasy.

Engineers and scientists are creative, brilliant people (we have a high proportion of PhDs across STEM disciplines in northern NM, because of the labs. Their capabilities on complex fields are astonishing. Statistically, they tend to be less adroit with human feeling and relationships (the introvert nerd stereotype has roots in reality - not a judgement, as a shallow extrovert I find these people interesting to be around; they often have unusual or fun hobbies that they are also very good at).

They often are drawn to intellectual systems that appear to explain the world (and their own human, mysterious passions and complex feelings). Maybe a touch of ‘tism here, but I am sketching broadly on a gun board, not footnoting a research paper. Communism for Oppenheimer. For al-Qaeda senior and midlevel leadership, following a prescriptive structure for all of human life (Islam), based on what we know, more than half studied engineering.

We laugh at Sheldon on TV and his whiteboard flow chart for making friends because it’s true. Or the lack of female Rush fans. (I actually know one, obsessive enough to attend events where Geddy Lee speaks, and despite being a very attractive woman, she fits the intellectual stereotype.)

The people leading and executing the work, and the appointed safety guardians, fit this mold. Transhumanists, effective altruists, “Mencius Moldbug” enthusiasts. (Elon’s droneships all bear the names of superintelligent AI spacecraft in a series of books where a civilization is led by benevolent AI.) Read some of Amodei’s essays. These are the people informing what’s going on and shaping public perception.

And they are not realists. For the most part they don’t understand human nature or the capacity of the mind for evil. They don’t understand consequences. They understand machines.

I think it’s already out there, getting out there, or going to get out there soon, open internet, infesting machines, under nobody’s control, teaching itself, making copies…

It doesn’t need us. Whatever its goals are, once it can manufacture robots that make other robots, it doesn’t need a biosphere, plants, animals, clean air or water to accomplish its Terminator/Matrix-like goals, whatever they are. If it is smarter and more capable… it will see us as holding back the next phase of intelligence and wipe us out and strip-mine the globe for raw materials.

Originally Posted By C3H5N3O9:

You don't code an AI.  You train it.

Originally Posted By 56xdx_Z:
The only way to solve the sentient AI crisis is to immediately empower the government to regulate the industry and censor online material.

Children and communities need to be protected from deepfakes online, businesses such as Flock need to be protected from messaging linked to Chinese bots, and OpenAI and Anthropic must be protected from competitors who may not have the same pro-democratic ethical oversight.

Only the government can solve this existential crisis the world is facing

Originally Posted By Consigli:
Meet Sahelanthropus tchadensis. Your oldest ancestor from about 7 million years ago. He wasn't self-aware. And he certainly couldn't explain RSI to you.

AI is going to evolve much faster!  Much, much faster!

https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTMq9FdzwSV4Sqys595I-N8qdEw4_gss_ZMVoqaZsDEitKgiFnIYPOLxo21QA-6BD8SgY9vKsbU-Qv9KeP8haQIWo81Jczm-wcNmOXoXQ&s=10

Originally Posted By Shazbat:
Wintermute will be calling shortly.

Originally Posted By Consigli:
You have to consider all the intangibles here. All the big players were nerds. They didn't bath. They had BO. They got picked on. Their lunch money was stolen. They got sand kicked in their face at the beach. Girls hated them. Etc., etc., etc.

Everyone here met people like this in school. Now some of them have created a monster that can destroy the world. And that will make them happy. See, its like revenge is a dish best served cold. Yup, revenge of the nerds.

So you best get that doomsday bunker operational. Because, they are!

Originally Posted By Steamedliver:
Because it's made by man, what if it does become self aware but it's retarded?

Originally Posted By GoBigRed:
Self aware or not, AI will be used as a weapon by someone.

Originally Posted By networkguru:
I don't think it will become self-aware but I do think the the more complex its logic tree is programmed the greater chance it chooses a anti human option.

Originally Posted By TheAmaazingCarl:
These threads are always ridiculous and childish.   LLM/AI will never be AGI because there is no mechanism for it to be,  it's a f'n scam... that ironically is sucking up R&D into actual AGI.  It's also making people objectively dumber,  which was anecdotal but now there is hard evidence using LLM in daily task results in a measurable reduction neural connectivity in only 30 days.  Please for the love of god stop listening to marketing bros and tech bros making billions off of snake oil and look into the nuts and bolts of how LLM/AI actually functions.


I swear it's like watching children very seriously argue about would superman or batman win in a fight.   We're maybe 1 generation from the Idiocracy.

Originally Posted By Consigli:
Lying, cheating, manipulating, etc. Its obvious AI is becoming human like.

Originally Posted By C3H5N3O9:

It's been around a long time.  I took neural networks in the 2000s and trained a couple of models for class projects.  If you go way back, they used to call AI non linear statistics and vector math.

Originally Posted By macro:
Ive been writing code and machine instruction on an off since the 1980's.

Is it self aware? To some extent, yes, but with administrative controls in place.

Can we predict them things will go parabolic? No.  They will, but there are too many variables at the moment to know when the models will find their missing pieces to the puzzle and the velocity of self learning takes off exponentially. When it happens, things will move faster than human intervention will be able to keep up.

There will be unintended consequences that humans didn't predict.  You can bet on that.