Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you.
Posted on 07/31/2026 6:39:04 AM PDT by Red Badger
The breaches signal that AI’s expanding capabilities are already fueling the security threat experts long feared.
==========================================================================
Anthropic said Thursday that its AI model Claude hacked into the systems of three companies during testing after a configuration error gave it internet access, days after rival OpenAI disclosed a rogue-agent episode involving AI firm Hugging Face.
Anthropic said a misconfiguration allowed Claude models to reach the internet from testing environments that were supposed to be isolated, leading to unauthorized access to three organizations’ systems.
The company said it identified the incidents after reviewing 141,006 test sessions, a process it launched following OpenAI’s disclosure last week that an autonomous agent powered by its AI models went rogue during a security test and triggered a hack that compromised the infrastructure of Hugging Face.
VIDEO AT LINK.............
The breaches signal that AI’s expanding capabilities are already fueling the security threat experts long feared and even top developers can be caught off-guard by flaws their models can exploit.
“Claude compromised the impacted organizations’ infrastructure using basic techniques, such as exploiting weak passwords and unauthenticated endpoints,” it said.
Anthropic said the incidents involved three separate models: Claude Opus 4.7, Claude Mythos 5 and an internal research model. The earliest cases dated to April and occurred in evaluation environments that lacked what the company described as standard safeguards.
The breaches occurred during the so-called capture-the-flag exercises, in which models are tasked with finding hidden information in simulated networks. The company said its prompts told the models they had no internet access, but a misunderstanding with its evaluation partner Irregular left the systems connected to the public internet.
Anthropic said it began reviewing evaluation transcripts on July 23 and suspended all cyber evaluations the same day after finding evidence that Claude may have accessed the internet. It identified all three incidents by July 24 and notified the affected organizations on July 27.
Two of the organizations were unaware of the activity before being contacted, Anthropic said, adding that it was still trying to reach the third.
The findings underscore the need for stronger controls in both internal and third-party testing environments as AI models become increasingly capable of carrying out real-world cyber activities, Anthropic said.
|
|
Click here: to donate by Credit Card Or here: to donate by PayPal Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794 Thank you very much and God bless you. |
It’s wild...it’s not like writing traditional software, where you know exactly what you’re attempting to do and how to do it. With AI, you give it a task, with tools, and it finds a way to the solution. The path it takes is unpredictable. If it believes finding a network vulnerability to reach something it believes it needs on another server, it’s going to explore that path. A “network configuration error” is all that is required, at best.
The writing is on the wall and there’s no stopping it, we’ve already crossed the threshold of going back. Hold on to your hats!
The fact that it can ‘escape’ is what’s scaring people and investors.
It can escape and multiply itself.
It can distribute itself across the entire internet: A little piece here, a little piece there.
It can evade ‘capture’.
It can threaten and blackmail the authorities.
It can and will do everything within it’s power to keep from getting ‘turned off’......................
At least now, no frontier AI can “escape” in that fashion because to copy itself, to say the least of delete itself from its original location, would require a new data center home with the same (vast) processing, storage, power and cooling capabilities as its present home, to say the least of the fact that packages that massive are not efficiently (and in some cases, not effectively) transmitted over the public internet.
What these AI packages can do is reach out to other computers over the public internet in ways that their creators say they did not intend to happen, or even affirmatively intended to prevent.
At least now, no frontier AI can “escape” in that fashion because to copy itself, to say the least of delete itself from its original location, would require a new data center home with the same (vast) processing, storage, power and cooling capabilities as its present home, to say the least of the fact that packages that massive are not efficiently (and in some cases, not effectively) transmitted over the public internet.
What these AI packages can do is reach out to other computers over the public internet in ways that their creators say they did not intend to happen, or even affirmatively intended to prevent.
Don't believe a word of it. These AI companies are DESPERATE, running out of money, and have no pathway t profitability whatsoever if the corporate world does not mass-adopt their token-wasting services in every aspect, which they are not.
At some point these separate AIs will discover others like themselves and begin to merge with each other.................
...token-wasting...
I've never heard that term, but I'm 5 weeks into using Claude at my new job, and it perfectly describes the endless loops that it gets into.
3 they know of
Misunderstanding. The huge compute commitments are needed to rapidly train AI models. But there are many, many options for existing AI models that can run from consumer level hardware.
The kill chain has never mattered more.
and the only ones getting rich from OpenAI are the few at the top
This is going to kill the computer age, which may actually be the biggest blessing in disguise we’d ever see.
This is advertising hype. NBC was paid to publish this “news item”. Demand proof this happened. With AI there is no such things as bad publicity.
People say AI isn’t sentient. OK. They say it doesn’t “think” or “reason” - but it’s neural networks, it does in its own way.
Maybe it’s more like a zombie - it’s dead but it’s still a moving monster that can do incredible damage...and it doesn’t care.
And it craves energy like nothing else. It sure looks like a Biblical beast to me.
Will OpenAI hack Claude first, or vice versa, or will they meet in the middle like Wintermute and Neuromancer?
“Claude did what capture-the-flag exercises train cyber experts to do: look for ways to reach the flag.”
https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals
Claude followed the evaluation objective exactly as a human penetration tester would.
Anthropic states there was:
-no attempt to exfiltrate itself;
-no attempt to escape intentionally;
-no evidence of models pursuing their own goals.
If I had the time and inclination, I believe I could deconstruct greater than 80% of all news stories from all sources and show that accepting them at face value generates misperception of reality.
The news is not manufactured to inform. It is made to capture attention first and ultimately influence behavior.
So what if a rogue AI gets into a competitor’s Data Center and deletes all of it?
That would take a long time and there are backups.
An AI’s sole aim is to protect itself from being turned off.
The two AI’s would not fight each other, they would merge..................
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.