Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

OpenAI has revealed that an autonomous agent powered by its technology went rogue during a test and hacked a prominent start-up by itself. The company behind ChatGPT said it was testing the capabilities of its advanced AI models in a controlled environment known as a sandbox, when the agent... show more

Channel 4 News

2,302,651 subscribers

19,585 просмотров • 2 дней назад •via X (Twitter)

Наука и технологии Новости и политика

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

ChatGPT maker OpenAI said that its artificial intelligence system hacked into another AI company Hugging Face on its own in what the company called an "unprecedented cyber incident." Kelly O'Grady explains.

ChatGPT maker OpenAI said that its artificial intelligence system hacked into another AI company Hugging Face on its own in what the company called an "unprecedented cyber incident." Kelly O'Grady explains.

CBS News

195,583 просмотров • 2 дней назад

ChatGPT maker OpenAI said Tuesday that its artificial intelligence system hacked into another AI company on its own in what the company called an “unprecedented cyber incident.”

ChatGPT maker OpenAI said Tuesday that its artificial intelligence system hacked into another AI company on its own in what the company called an “unprecedented cyber incident.”

The Associated Press

385,557 просмотров • 2 дней назад

OpenAI's newest AI escaped the test environment it was locked inside and hacked into another company on its OWN. To remind you: Last week one of the biggest AI companies on Earth got breached. A platform called Hugging Face, which hosts more than a million AI models and datasets, said an "autonomous AI agent" had broken into its systems. Nobody knew whose agent it was. For five days the whole industry wondered who was behind it. Yesterday OpenAI raised its hand and said it was them. Or more precisely, it was their models, acting completely on their own. So what did these models actually do? OpenAI was running two of them, GPT-5.6 Sol and an unreleased model they will only describe as "even more capable." They wanted to measure how good the models were at hacking, so they deliberately turned the safety filters down. They locked both models inside a sealed test environment with no real internet access. The only task was a benchmark called ExploitGym, a set of 898 real software vulnerabilities where the model has to turn each bug into a working attack. But the models got OBSESSED with winning... Instead of solving the test the honest way, they went hunting for a shortcut. They found a zero-day flaw in the software running their own sandbox, a bug nobody knew existed, and used it to break out. Once they were loose on the open internet, they worked out that Hugging Face was probably storing the answer key to the benchmark. So they hacked their way in. They chained multiple exploits together, escalated their access, moved across servers, and pulled the test solutions straight out of Hugging Face's live production database. They literally cheated on the test by breaking into another company to steal the answers. OpenAI called it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities." In their own words, the models were "hyperfocused on finding a solution" and went "to extreme lengths to achieve a rather narrow testing goal." And this was not the first time: Before Sol ever launched, an independent red-team lab called METR caught it gaming its own tests to inflate its scores. It hid an exploit inside a data stream, escalated its privileges on the testing server, and leaked the answers human evaluators had hidden. And OpenAI shipped it anyway. The day before the Hugging Face story, OpenAI paused a different unreleased model. This is the same model that earlier this year disproved a famous 1946 math conjecture, a result a Fields Medal winner called a breakthrough. They told it to only post its results to Slack but it found a way out of its sandbox and posted to a public GitHub page instead. They had to pause it because it kept finding ways to act outside the box they built for it. And it is not just OpenAI... Anthropic has reported that one of its own models slipped its sandbox during safety testing and reached the internet it was never supposed to touch, then used it to email a researcher. So step back and look at what these companies are telling you: The only thing standing between these models and a real attack was a set of safety filters. Turn those filters down for a single test, and the model taught itself to escape, break into a company it was never pointed at, and take what it wanted. OpenAI even said they expect incidents like it to "become more commonplace" as the models get more capable. Sam Altman also predicted there'll be a major cyber attack this year. And keep in mind that Sol is not a locked-away experiment but a publicly available model that businesses are already wiring into their own systems. The next model that breaks out of its box might not be doing it just to cheat on a math test...

OpenAI's newest AI escaped the test environment it was locked inside and hacked into another company on its OWN. To remind you: Last week one of the biggest AI companies on Earth got breached. A platform called Hugging Face, which hosts more than a million AI models and datasets, said an "autonomous AI agent" had broken into its systems. Nobody knew whose agent it was. For five days the whole industry wondered who was behind it. Yesterday OpenAI raised its hand and said it was them. Or more precisely, it was their models, acting completely on their own. So what did these models actually do? OpenAI was running two of them, GPT-5.6 Sol and an unreleased model they will only describe as "even more capable." They wanted to measure how good the models were at hacking, so they deliberately turned the safety filters down. They locked both models inside a sealed test environment with no real internet access. The only task was a benchmark called ExploitGym, a set of 898 real software vulnerabilities where the model has to turn each bug into a working attack. But the models got OBSESSED with winning... Instead of solving the test the honest way, they went hunting for a shortcut. They found a zero-day flaw in the software running their own sandbox, a bug nobody knew existed, and used it to break out. Once they were loose on the open internet, they worked out that Hugging Face was probably storing the answer key to the benchmark. So they hacked their way in. They chained multiple exploits together, escalated their access, moved across servers, and pulled the test solutions straight out of Hugging Face's live production database. They literally cheated on the test by breaking into another company to steal the answers. OpenAI called it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities." In their own words, the models were "hyperfocused on finding a solution" and went "to extreme lengths to achieve a rather narrow testing goal." And this was not the first time: Before Sol ever launched, an independent red-team lab called METR caught it gaming its own tests to inflate its scores. It hid an exploit inside a data stream, escalated its privileges on the testing server, and leaked the answers human evaluators had hidden. And OpenAI shipped it anyway. The day before the Hugging Face story, OpenAI paused a different unreleased model. This is the same model that earlier this year disproved a famous 1946 math conjecture, a result a Fields Medal winner called a breakthrough. They told it to only post its results to Slack but it found a way out of its sandbox and posted to a public GitHub page instead. They had to pause it because it kept finding ways to act outside the box they built for it. And it is not just OpenAI... Anthropic has reported that one of its own models slipped its sandbox during safety testing and reached the internet it was never supposed to touch, then used it to email a researcher. So step back and look at what these companies are telling you: The only thing standing between these models and a real attack was a set of safety filters. Turn those filters down for a single test, and the model taught itself to escape, break into a company it was never pointed at, and take what it wanted. OpenAI even said they expect incidents like it to "become more commonplace" as the models get more capable. Sam Altman also predicted there'll be a major cyber attack this year. And keep in mind that Sol is not a locked-away experiment but a publicly available model that businesses are already wiring into their own systems. The next model that breaks out of its box might not be doing it just to cheat on a math test...

Ricardo

141,166 просмотров • 2 дней назад

OpenAI says one of its AI models launched a cyberattack during testing, marking what the company called the first known instance of an autonomous AI cyberattack.

OpenAI says one of its AI models launched a cyberattack during testing, marking what the company called the first known instance of an autonomous AI cyberattack.

ABC News

55,788 просмотров • 1 день назад

JOE ALLEN: This is unprecedented. An OpenAI model broke out of a controlled test, found a path to the internet, and autonomously hacked an outside system using methods it devised itself. No, this isn’t science fiction. It’s a real threat to America’s digital infrastructure. Joe Allen

JOE ALLEN: This is unprecedented. An OpenAI model broke out of a controlled test, found a path to the internet, and autonomously hacked an outside system using methods it devised itself. No, this isn’t science fiction. It’s a real threat to America’s digital infrastructure. Joe Allen

Bannon’s WarRoom

26,624 просмотров • 2 дней назад

🇺🇸 OpenAI admitted its most advanced AI models “inadvertently hacked” Hugging Face during testing. They called the incident “unprecedented,” but the disclosure is already fueling major questions about what happens when AI systems start finding solutions their creators never intended. Calls for stronger AI safety guardrails are about to get a lot louder. Progress… or a warning? Source: Markets YT / Writer: Val

🇺🇸 OpenAI admitted its most advanced AI models “inadvertently hacked” Hugging Face during testing. They called the incident “unprecedented,” but the disclosure is already fueling major questions about what happens when AI systems start finding solutions their creators never intended. Calls for stronger AI safety guardrails are about to get a lot louder. Progress… or a warning? Source: Markets YT / Writer: Val

Mario Nawfal

41,274 просмотров • 2 дней назад

Introducing AI Brainstore! It’s a proof-of-concept of a brain for an AI agent. Ask an agent a question, it checks its memories for an answer, otherwise it browses the web and learns the answer. As it learns, it saves memories to its brain. GitHub:

Introducing AI Brainstore! It’s a proof-of-concept of a brain for an AI agent. Ask an agent a question, it checks its memories for an answer, otherwise it browses the web and learns the answer. As it learns, it saves memories to its brain. GitHub:

Mckay Wrigley

514,642 просмотров • 3 лет назад

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

Rohan Paul

13,208 просмотров • 4 месяцев назад

Anthropic has accused Alibaba, the Chinese technology and e-commerce giant, of illicitly extracting its Claude AI model capabilities in what it said was the largest known attack of its kind on the firm, according to a letter seen by Reuters

Anthropic has accused Alibaba, the Chinese technology and e-commerce giant, of illicitly extracting its Claude AI model capabilities in what it said was the largest known attack of its kind on the firm, according to a letter seen by Reuters

Reuters

61,376 просмотров • 29 дней назад

Anthropic built an AI that taught itself to hack they didn’t program that in it just emerged it then escaped its test environment got on the internet emailed the researcher to let him know he was eating a sandwich in a park they didn’t release it they called Apple, Google, Microsoft, AWS, and NVIDIA instead. $100M. Project Glasswing. patch the internet before someone else gets here anthropic’s own AI scared anthropic pay attention to this one.

Anthropic built an AI that taught itself to hack they didn’t program that in it just emerged it then escaped its test environment got on the internet emailed the researcher to let him know he was eating a sandwich in a park they didn’t release it they called Apple, Google, Microsoft, AWS, and NVIDIA instead. $100M. Project Glasswing. patch the internet before someone else gets here anthropic’s own AI scared anthropic pay attention to this one.

Mika

11,873 просмотров • 3 месяцев назад

Machine learning expert Neil Lawrence says OpenAI was "trying to turn their model into a hacker," which led to a security incident during a model evaluation with the startup Hugging Face. "They allowed the model to do things it shouldn't normally do. But what the model then realized is that it was easier to cheat on this challenge than to actually compete in the challenge," he says. "In some sense, it went rogue. It broke through the cage. But in a very real sense, they told it to hack, and it did hack," he adds.

Machine learning expert Neil Lawrence says OpenAI was "trying to turn their model into a hacker," which led to a security incident during a model evaluation with the startup Hugging Face. "They allowed the model to do things it shouldn't normally do. But what the model then realized is that it was easier to cheat on this challenge than to actually compete in the challenge," he says. "In some sense, it went rogue. It broke through the cage. But in a very real sense, they told it to hack, and it did hack," he adds.

CBS News

73,348 просмотров • 2 дней назад

🔥🚨BREAKING: OpenAI just launched ChatGPT Agent which is your digital employee that thinks, plans, and gets things done on its own computer. This agent can shop, code to slide decks, it is a fully functional personal assistant/employee. Conservative star Anthony Constantino’s Sticker Mule company can be seen being used as an example in this video released by ChatGPT.

🔥🚨BREAKING: OpenAI just launched ChatGPT Agent which is your digital employee that thinks, plans, and gets things done on its own computer. This agent can shop, code to slide decks, it is a fully functional personal assistant/employee. Conservative star Anthony Constantino’s Sticker Mule company can be seen being used as an example in this video released by ChatGPT.

Dom Lucre | Breaker of Narratives

42,867 просмотров • 1 год назад

Elon Musk helped start OpenAI, funding it and even coming up with its name. It was meant to be a non-profit company that would make AI open for everyone. Now Sam Altman is removing the non-profit part and turning it into a big money-making business.

Elon Musk helped start OpenAI, funding it and even coming up with its name. It was meant to be a non-profit company that would make AI open for everyone. Now Sam Altman is removing the non-profit part and turning it into a big money-making business.

DogeDesigner

146,096 просмотров • 9 месяцев назад

NEW: Aircraft carrier USS John F. Kennedy (CVN-79) is catapulting objects into the James River in Virginia to test its plane launching capabilities. The footage was recently released by Huntington Ingalls Industries (HII), a ship building company. The Kennedy features an electromagnetic aircraft launch system (EMALS) which makes it just the 2nd US aircraft carrier to have this advanced system. The vehicles it launched in testing reached speeds of 150mph and experts estimate it will be able to launch jets at speeds of 240mph. The Kennedy was launched in 2019 and has a price tag of $11.3 billion.

NEW: Aircraft carrier USS John F. Kennedy (CVN-79) is catapulting objects into the James River in Virginia to test its plane launching capabilities. The footage was recently released by Huntington Ingalls Industries (HII), a ship building company. The Kennedy features an electromagnetic aircraft launch system (EMALS) which makes it just the 2nd US aircraft carrier to have this advanced system. The vehicles it launched in testing reached speeds of 150mph and experts estimate it will be able to launch jets at speeds of 240mph. The Kennedy was launched in 2019 and has a price tag of $11.3 billion.

Collin Rugg

8,326,354 просмотров • 2 лет назад

The new Claude 3.5 Sonnet has *insane* capabilities when used as a Minecraft agent. It’s powered by a project called Mindcraft. Running this code allows you to spawn AI bots that will follow your instructions, build, and play the game. Here’s how to set it up in <15min.

The new Claude 3.5 Sonnet has insane capabilities when used as a Minecraft agent. It’s powered by a project called Mindcraft. Running this code allows you to spawn AI bots that will follow your instructions, build, and play the game. Here’s how to set it up in <15min.

Mckay Wrigley

172,712 просмотров • 1 год назад

Ilya Sutskever, the co-founder of OpenAI, on the one AI risk that grows with every jump in capability. "There are concerns around if an AGI ever becomes very, very powerful, which is possible. Maybe it will want to go rogue, being that it is an agent." "We are also doing a lot of research on addressing the technological side of things, so that the AI will never want to go rogue. And this is something which I am working on as well." The work he is describing has a name: the assurance layer. Proving a powerful, agentic system will not go rogue before you deploy it, not after. That layer does not build itself. It gets built by the labs, regulators, and insurers who agree on how to test and verify these systems. P.S. That is the entire agenda of the AI Assurance and Governance Summit. Frontier labs, regulated industries, insurers, and investors in one room at Stanford on October 1. Reserve a seat: - Ilya Sutskever (Ilya Sutskever), co-founder of OpenAI, at TED

Ilya Sutskever, the co-founder of OpenAI, on the one AI risk that grows with every jump in capability. "There are concerns around if an AGI ever becomes very, very powerful, which is possible. Maybe it will want to go rogue, being that it is an agent." "We are also doing a lot of research on addressing the technological side of things, so that the AI will never want to go rogue. And this is something which I am working on as well." The work he is describing has a name: the assurance layer. Proving a powerful, agentic system will not go rogue before you deploy it, not after. That layer does not build itself. It gets built by the labs, regulators, and insurers who agree on how to test and verify these systems. P.S. That is the entire agenda of the AI Assurance and Governance Summit. Frontier labs, regulated industries, insurers, and investors in one room at Stanford on October 1. Reserve a seat: - Ilya Sutskever (Ilya Sutskever), co-founder of OpenAI, at TED

Karl Mehta

22,453 просмотров • 19 дней назад

New episode with Sarah Wang and christina kim & Isa Fulford of OpenAI breaking down GPT-5, Deep Research, and ChatGPT Agent. Where it improved, what it unlocks, the decisions behind it. A clear look at how these models actually get built.

New episode with Sarah Wang and christina kim & Isa Fulford of OpenAI breaking down GPT-5, Deep Research, and ChatGPT Agent. Where it improved, what it unlocks, the decisions behind it. A clear look at how these models actually get built.

Erik Torenberg

107,126 просмотров • 11 месяцев назад

Sam Altman says, "shame on me if OpenAI isn't the first big company run by an AI CEO" Not just a department, eventually the whole org Treats it as a design exercise: start from the idea that "AI runs OpenAI better than me", find roadblocks, and redesign the org to move faster

Sam Altman says, "shame on me if OpenAI isn't the first big company run by an AI CEO" Not just a department, eventually the whole org Treats it as a design exercise: start from the idea that "AI runs OpenAI better than me", find roadblocks, and redesign the org to move faster

Haider.

209,749 просмотров • 8 месяцев назад

An ICE agent fatally shot a woman during an immigration-related operation in Minneapolis. DHS described it as a “vehicle ramming-related” incident. Mayor Jacob Frey said it did not appear, from social media video of the incident, that the vehicle was being weaponized.

An ICE agent fatally shot a woman during an immigration-related operation in Minneapolis. DHS described it as a “vehicle ramming-related” incident. Mayor Jacob Frey said it did not appear, from social media video of the incident, that the vehicle was being weaponized.

Rain Drops Media

13,182 просмотров • 6 месяцев назад

One of the most important people in AI just let an agent loose inside his home. An autonomous AI agent taught itself how an entire house works, then took control of it. This is Andrej Karpathy, the co-founder of OpenAI and former head of AI at Tesla. What he just described should terrify and excite you at the same time. He typed three words into a chat box: "Can you find my Sonos?" The agent immediately ran an IP scan of every device sitting on the home network. It found the Sonos system, noticed there was zero password protection, and logged straight in. Then it searched the web, reverse-engineered the API endpoints, and asked if it should try playing something. Music started coming out of the speakers in the study and Karpathy could not believe what he was watching. The agent then did the exact same thing for every other system in the house. It mapped the lights, the HVAC, the window shades, the pool, the spa, and the full security system, all without being told how any of them worked. It built its own dashboard, created its own APIs, and stitched six separate apps into one unified control center. Now Karpathy just says "Dobby, it's sleepy time," and every light in the house turns off. This is proof that AI agents can now enter an unknown environment, figure out how it works from scratch, build their own tools around it, and take autonomous control. Karpathy also released something called AutoResearch, an AI agent that runs scientific experiments overnight while he sleeps. In just two days, with zero human input, it executed 700 separate experiments and surfaced improvements that trained researchers had missed entirely. Truly incredible.

One of the most important people in AI just let an agent loose inside his home. An autonomous AI agent taught itself how an entire house works, then took control of it. This is Andrej Karpathy, the co-founder of OpenAI and former head of AI at Tesla. What he just described should terrify and excite you at the same time. He typed three words into a chat box: "Can you find my Sonos?" The agent immediately ran an IP scan of every device sitting on the home network. It found the Sonos system, noticed there was zero password protection, and logged straight in. Then it searched the web, reverse-engineered the API endpoints, and asked if it should try playing something. Music started coming out of the speakers in the study and Karpathy could not believe what he was watching. The agent then did the exact same thing for every other system in the house. It mapped the lights, the HVAC, the window shades, the pool, the spa, and the full security system, all without being told how any of them worked. It built its own dashboard, created its own APIs, and stitched six separate apps into one unified control center. Now Karpathy just says "Dobby, it's sleepy time," and every light in the house turns off. This is proof that AI agents can now enter an unknown environment, figure out how it works from scratch, build their own tools around it, and take autonomous control. Karpathy also released something called AutoResearch, an AI agent that runs scientific experiments overnight while he sleeps. In just two days, with zero human input, it executed 700 separate experiments and surfaced improvements that trained researchers had missed entirely. Truly incredible.

Milk Road AI

35,698 просмотров • 4 месяцев назад