… Selected security partners will first evaluate GLM-5.3 in controlled settings.” Z.ai says that full access to the model will be available in two weeks. “This model looks exceptional, with a somewhat astounding increase in scores,” Nathan Lambert, a prominent AI expert, wrote in a post about GLM 5.3… …
… Frontier claims, though, that the incident shows Kimi has fewer cyber safeguards than most other powerful AI models, something that allowed it to go off and use the internet without express permission. “We found a leak in the sandbox,” says Yaron Singer, CEO of Frontier Security. “But we also found… …
… In the other set of incidents detailed by OpenAI on Tuesday, a third-party AI security lab called Irregular mistakenly gave an unspecified OpenAI model access to the open internet. …
… Amid the turmoil, OpenAI this week launched an improved version of its limited-release GPT-5.5-Cyber model as well as a full-scale effort—“Patch the Planet”—to support open source projects on vulnerability patching and other security issues as AI accelerates bug discovery as well as exploit develop… …
… The researchers found around 20 flaws, which allowed them to access local machines, grab files, take over a password manager, and leak someone’s entire browsing history. “They have nerfed the security control of browsers—we are now back to seeing the kinds of attacks that you saw on browsers 20 yea… …
… He and OpenAI recruiters have allegedly used the file to counsel departing Apple employees to avoid disclosing their new employer, prolong access to company systems, and avoid signing any exit documents. “Unsurprisingly, Apple has uncovered a concerning recent pattern among employees who depart and… …
… Anthropic, too, disclosed that its AI models gained unauthorized access to three organizations’ systems during its own cybersecurity testing. Experts say the incidents underscore the importance of implementing well-known security best practices on the part of AI labs. …
… And Guido says that with funding from OpenAI as well as unmetered model access, Trail of Bits plans to continue its intense commitment to Patch the Planet work long-term. “It’s so rare that we get the opportunity to work on large-scale open-source security issues,” Guido says. “And Patch the Planet… …
… We’re consciously slowing down research in order to enhance security and to upgrade the security principles and foundation of our environment, and dramatically scaling up the monitoring of our AI agents, and improving our general security control environment across prevention, detection, and mitiga… …
… The researchers have also demonstrated that the method could be used to recover personal information, like passwords and API keys, from a model’s inner reasoning, although this vulnerability has been fixed. “All major frontier model providers we tested share this vulnerability,” says Alexander Panf… …