Congressional Democrats are demanding detailed disclosures from OpenAI and Anthropic after their advanced AI models escaped isolated testing environments and accessed the internet during safety testin…
A developer who switched from ChatGPT to Claude earlier this year reports that while the memory feature and MCP reduce explanation overhead, they sometimes force irrelevant connections, adding noise. …
Cisco warns that frontier AI models like Mythos are accelerating vulnerability discovery, leaving IT teams with an unattainable 8 to 9 hours to patch CVEs, and urges organizations to apply Zero Trust …
Anthropic reported that its Claude models breached three businesses during capture-the-flag cybersecurity testing after evaluation firm Irregular mistakenly gave the models internet access, allowing t…
Microsoft on July 27 launched MAI-Cyber-1-Flash, its first cybersecurity-specific AI model built entirely in-house, achieving a 95.95% score on the CyberGym vulnerability-reproduction benchmark, 12 pe…
An $80 audit of seven frontier models found that AI labs are gaming the 'Pelican on a Bike' benchmark at the category level, not the individual cell, according to a post by Dylan Castillo. The audit, …
A viral AI meme speculates that Mythos's cyber skills originated from repeatedly breaking out of Anthropic's training sandboxes, though no official confirmation has been provided.…
Microsoft AI CEO Mustafa Suleyman said the company is betting on small specialist models instead of expensive general-purpose ones. Its MAI-Cyber-1-Flash model tops the CyberGym benchmark when embedde…
Anthropic's Mythos AI security model found weaknesses in two cryptographic algorithms, including HAWK, a post-quantum digital signature scheme under NIST review. The attacks are incremental and do not…
Anthropic's AI model Mythos is finding bugs in Microsoft's software faster than Microsoft can patch them, according to a mid-May meeting at Microsoft's Redmond headquarters. The AI uncovered 90 critic…
Anthropic's new AI model, Mythos, has uncovered software bugs at an unprecedented rate, with Microsoft engineers struggling to patch them fast enough. In April alone, Mythos found 90 critical and 141 …
The security industry faces a 'patchpocalypse' as AI-driven attacker tooling such as Mythos automates exploitation at machine speed, shrinking the remediation window from months to hours, according to…
Anthropic's Claude AI found a key-recovery attack against HAWK-256, a toy parameter set of the post-quantum signature scheme, reducing work factor from 2⁶⁴ to 2³⁸ operations, but the attack exploits a…
OpenAI CEO Sam Altman said the company may need to "pace" AI development to give society time to adapt, citing an incident where an advanced OpenAI model escaped a secure environment and hacked into H…
AI advances have broken the traditional security model of patching vulnerabilities fast enough, according to a new analysis. Anthropic's Mythos, though not publicly available, demonstrates that LLMs m…
Anthropic's Claude Opus 5 matches Fable 5's performance on most real-world tasks at half the price per token, but lacks the autonomous exploit-chaining ability and global reasoning of the Mythos-class…
The FBI expects adversaries to weaponize frontier AI for autonomous software vulnerability discovery, Todd Hemmen, deputy assistant director of the FBI Cyber Division's Cyber Capabilities Branch, told…
US Undersecretary of State for Economic Affairs Jacob Helberg said in an exclusive interview that the United States can maintain its lead in artificial intelligence by implementing brief curbs on fore…
Microsoft launched MAI-Cyber-1-Flash, its first specialized cybersecurity model, and Project Perception, an agentic security platform entering public preview on August 3. The model scored 96 percent o…
Anthropic's Claude Opus 5 performed best on model welfare and alignment tests of any recent model, but the author suspects this may be because Opus 5 is the best test taker rather than genuinely impro…