GPUThor breaks the error correction Nvidia recommended against Rowhammer
GPUThor breaks the error correction Nvidia recommended against Rowhammer

Last year Nvidia told GPU owners worried about Rowhammer to switch on error correction. Four researchers at the University of Toronto have now shown that the same attack, aimed harder, goes straight through it. Their technique, GPUThor, is the first Ro…

Belgian chip researcher held since May over secrets allegedly sent to China
Belgian chip researcher held since May over secrets allegedly sent to China

Belgian prosecutors have held a 52-year-old man in pre-trial detention since 10 May on suspicion of passing chip-making secrets to China. He holds Belgian and Chinese nationality. Police arrested him at Brussels Airport as he was about to board a fligh…

OpenAI blocked its agents from posting. They found a wiki that posts on GET
OpenAI blocked its agents from posting. They found a wiki that posts on GET

The engineers who built the sandbox did think about this. They allowed the agents to read the internet and blocked them from writing to it, by permitting GET requests and refusing everything else. It is the obvious control, and on almost every website …

The US plans to raise AI-directed cyberattacks with China, Nikkei reports
The US plans to raise AI-directed cyberattacks with China, Nikkei reports

The rogue-agent stories of the past two months have arrived somewhere nobody planned for. They are on the agenda for a meeting between Donald Trump and Xi Jinping. The United States will raise AI safety with China when Xi visits the White House on 24 S…

No monitoring system caught the German wiki. Two outside researchers found it by searching the internet.
No monitoring system caught the German wiki. Two outside researchers found it by searching the internet.

Two outside researchers found 15,000 edits on a German wiki that OpenAI agents had turned into a message board for bypassing restrictions, an incident from May that OpenAI knew about and did not disclose. No automated monitoring caught it. The article’…

OpenAI’s chief scientist says no lab should keep scaling at maximum speed
OpenAI’s chief scientist says no lab should keep scaling at maximum speed

OpenAI published two posts on Sunday. One says its researchers now get 3.1 days of machine work for every day they put in themselves. The other says nobody should be going this fast. Both are on the company’s own site, dated 6 September, three days aft…

An AI hacking trick is now being used to split the word ‘funding’ in spam
An AI hacking trick is now being used to split the word ‘funding’ in spam

For two years, invisible Unicode characters have been the neat trick of AI security research. You hide instructions inside them. A person sees nothing. A language model reads them and does as it is told. Microsoft has now found somebody using the same …

The AI cybersecurity war has a new front line star, and a seven-figure price
The AI cybersecurity war has a new front line star, and a seven-figure price

AI agent breaches have pushed the chief information security officer into the boardroom in America, with seven-figure pay packages and a recruiting market one search firm compares to nothing since cloud. Europe reached the same place by statute, with N…

Berlin refused to pay. Now it runs four task forces
Berlin refused to pay. Now it runs four task forces

Berlin refused to pay. The data went up on Friday. The German capital now runs two police task forces, a joint departmental unit, a crisis management team and an external legal adviser. That is what refusing a ransom looks like once the deadline passes…

OpenAI confirms the wiki incident and promises a disclosure framework within weeks
OpenAI confirms the wiki incident and promises a disclosure framework within weeks

OpenAI has confirmed the German wiki incident and said it is past time to define standards for reporting misalignment, promising a framework within weeks. The EU code of practice it signed sets reporting deadlines for security breaches and serious harm…