| Mobile Web

Search results for Trusted Access for Cyber

AI & Enterprise

Opus 4.7 poses sufficient threat; government to unveil Mythos response measures as early as late May

The government plans to announce a comprehensive response to AI-driven cybersecurity threats as early as late May, after a case was found in which Anthropic’s general-purpose AI model Opus 4.7 successfully penetrated a corporate service. Officials said the test, conducted with a company’s consent, identified weaknesses in minutes. Authorities have asked thousands of firms to strengthen security and issued guidance for AI-based cyberattacks.

AI & Enterprise

OpenAI expands TAC cybersecurity program, offers GPT-5.5 Cyber preview

OpenAI said on Friday it is expanding its Trusted Access for Cyber (TAC) program for security professionals and will provide participants with a preview of GPT-5.5 Cyber. TAC has three tiers and requires identity verification, with eligible participants including corporate security teams, bug bounty hunters, verified researchers and some government agencies. OpenAI said thousands of verified defenders and hundreds of teams are currently participating.

AI & Enterprise

South Korea science ministry discusses AI security response with industry, academia on Mythos

South Korea\'s Ministry of Science and ICT held a meeting on Friday with experts from industry, academia and research institutes to discuss the domestic impact of cybersecurity projects by global AI firms including Anthropic and OpenAI and possible responses. The ministry previously asked about 30,000 companies\' chief information security officers to check their security readiness and later distributed additional guidance on AI-based cyberattack preparedness and CEO actions.