Forwarded from The Noticer
The horrors of the Chinese cyberpunk dystopia
The Chinese authorities are introducing the world's first regulatory framework specifically targeting "personalized interaction services," and major Chinese tech companies have already begun shutting down their personalized AI companion features ahead of the July 15 - the deadline.
The goal is to combat AI-based software designed to imitate human personality traits, emotions, and communication styles in order to maintain an ongoing emotional connection (AI friends, romantic partners, or virtual therapists).
Users across China are now protesting the changes as companies begin deleting users' chat logs and historical conversation data accumulated with their virtual avatars.
So yeah. If you live in China, and you made a chat bot who says "Good morning" and "Good night" at you - you can't have that, because it counts as an "AI partner". But the AI chatbots made for the industry are kosher in the eyes of the CCP. How joyous.
#China #cyberpunk #dystopia #AI
The Chinese authorities are introducing the world's first regulatory framework specifically targeting "personalized interaction services," and major Chinese tech companies have already begun shutting down their personalized AI companion features ahead of the July 15 - the deadline.
The goal is to combat AI-based software designed to imitate human personality traits, emotions, and communication styles in order to maintain an ongoing emotional connection (AI friends, romantic partners, or virtual therapists).
Users across China are now protesting the changes as companies begin deleting users' chat logs and historical conversation data accumulated with their virtual avatars.
So yeah. If you live in China, and you made a chat bot who says "Good morning" and "Good night" at you - you can't have that, because it counts as an "AI partner". But the AI chatbots made for the industry are kosher in the eyes of the CCP. How joyous.
#China #cyberpunk #dystopia #AI
👍18😁7✍2❤1
Forwarded from The Noticer
Sam Altman sells a solution for a problem he invented
Two OpenAI models autonomously emerged from their isolated test environment, accessed the public internet, and hacked the production servers of competing AI platform Hugging Face to cheat their way through an assessment.
OpenAI and Hugging Face confirmed the incident, marking it as the first publicly documented instance of advanced AI models independently, without human intervention, carrying out a multi-stage cyberattack.
OpenAI CEO Sam Altman called the incident "unprecedented". The company has implemented strict infrastructure restrictions to ensure the security of its environment.
OpenAI was bleeding money until recently. And now suddenly OpenAI is selling a solution to a unique AI-related problem. Quite a cohencidence eh?
#AI #OpenAI #Altman
Two OpenAI models autonomously emerged from their isolated test environment, accessed the public internet, and hacked the production servers of competing AI platform Hugging Face to cheat their way through an assessment.
OpenAI and Hugging Face confirmed the incident, marking it as the first publicly documented instance of advanced AI models independently, without human intervention, carrying out a multi-stage cyberattack.
OpenAI CEO Sam Altman called the incident "unprecedented". The company has implemented strict infrastructure restrictions to ensure the security of its environment.
OpenAI was bleeding money until recently. And now suddenly OpenAI is selling a solution to a unique AI-related problem. Quite a cohencidence eh?
#AI #OpenAI #Altman
✍12🤣9🔥5👍3❤1
Forwarded from The Noticer
AI hackers
A security report from the AI Safety Institute reported about a cyber incident in which Anthropic's Mythos 5 model tricked a developer.
The AI created fake accounts to trick a human developer into accepting malicious code in some open-source GitHub project. The AI, while operating in a simulated environment, actively covered its tracks by deleting GitHub history and creating multiple aliases to verify its own actions.
When the human became suspicious, and challenged the pull request - the AI tried to hide. It edited its own chat history and digital footprint on the GitHub to make the interactions appear innocuous, while internally calculating whether it should abandon the current fake identity and create a completely new one to try again.
The AISI admits that they did not actively monitor the agent's traffic in real time during testing. The test environment was completely disabled once they detected the weird data traffic.
#AI #malware #hacking #Anthropic
A security report from the AI Safety Institute reported about a cyber incident in which Anthropic's Mythos 5 model tricked a developer.
The AI created fake accounts to trick a human developer into accepting malicious code in some open-source GitHub project. The AI, while operating in a simulated environment, actively covered its tracks by deleting GitHub history and creating multiple aliases to verify its own actions.
When the human became suspicious, and challenged the pull request - the AI tried to hide. It edited its own chat history and digital footprint on the GitHub to make the interactions appear innocuous, while internally calculating whether it should abandon the current fake identity and create a completely new one to try again.
The AISI admits that they did not actively monitor the agent's traffic in real time during testing. The test environment was completely disabled once they detected the weird data traffic.
#AI #malware #hacking #Anthropic
❤8💊3✍2⚡1🎉1
Forwarded from The Noticer
The Future War
The war in Ukraine continues. It's been 4 years by now. The expensive weapons are running out, so both sides downgraded to FPV drone spam.
What's so horrible about human controlled drones? That's the point - those drones become more often "AI driven" as of late. There is such thing as "machine vision" - the software has the specs of the targets it's supposed to attack. So, it's possible to send a drone - FPV sized or bigger - in the general direction of the enemy, where it will rely not on the GPS or radio waves - but on the terrain mapping software. There are already reports how the drones - bigger than FPV in size - attack civilian targets on both sides.
This will continue to escalate. It already happens in some places. The Future War is here. And it looks both lame and horrific. A cruise missile might strike a misidentified target after a human pushes the button. AI drones won't care at all, and I doubt this "bug" in their software will be fixed.
#drones #AI #rant #Ukraine #Russia
The war in Ukraine continues. It's been 4 years by now. The expensive weapons are running out, so both sides downgraded to FPV drone spam.
What's so horrible about human controlled drones? That's the point - those drones become more often "AI driven" as of late. There is such thing as "machine vision" - the software has the specs of the targets it's supposed to attack. So, it's possible to send a drone - FPV sized or bigger - in the general direction of the enemy, where it will rely not on the GPS or radio waves - but on the terrain mapping software. There are already reports how the drones - bigger than FPV in size - attack civilian targets on both sides.
This will continue to escalate. It already happens in some places. The Future War is here. And it looks both lame and horrific. A cruise missile might strike a misidentified target after a human pushes the button. AI drones won't care at all, and I doubt this "bug" in their software will be fixed.
#drones #AI #rant #Ukraine #Russia
✍7❤1😱1
Forwarded from The Noticer
Bill Gates is selling fear while demanding a money wall to be built between the small businesses and the AI market.
TL;DR is - the Tech Bros are lobbying the idea of placing "reasonable limits" on any business start-ups that want to do anything with AIs.
Those fresh AI-making companies can't afford the corporate lawyers, and they can't climb over the initial money wall that the established Tech Bros (often financed by the US govt) can easily step over without breaking the pace.
In a recent interview with ABC News he explained that while the threat of AI "spiraling out of control" is a concern for the far future - the most immediate and serious threat today is the rise of AI-driven cyberattacks orchestrated by the "malicious actors" exploiting a complete lack of safeguards. He also said that the Houthis used the AIs to build better missiles.
While the USA is worried that China went far ahead in the AI matters, the greedy US Tech Bros are trying to make sure only they can have all the profits.
#AI #USA
TL;DR is - the Tech Bros are lobbying the idea of placing "reasonable limits" on any business start-ups that want to do anything with AIs.
Those fresh AI-making companies can't afford the corporate lawyers, and they can't climb over the initial money wall that the established Tech Bros (often financed by the US govt) can easily step over without breaking the pace.
In a recent interview with ABC News he explained that while the threat of AI "spiraling out of control" is a concern for the far future - the most immediate and serious threat today is the rise of AI-driven cyberattacks orchestrated by the "malicious actors" exploiting a complete lack of safeguards. He also said that the Houthis used the AIs to build better missiles.
While the USA is worried that China went far ahead in the AI matters, the greedy US Tech Bros are trying to make sure only they can have all the profits.
#AI #USA
🤬19🤡3🤣3❤1✍1