UK Startup HarmonicSecurity Secures $7 Million in Funding


OpenAI has launched GPT-5.5-Cyber, its advanced cybersecurity AI model, to EU enterprises, government agencies, and cybersecurity experts, marking a key step in global cyber defense collaboration and raising concerns over tech giants' security compliance.....
Google's Threat Intelligence team first discovered and intercepted an AI-developed zero-day vulnerability attack, indicating that AI has evolved from a supportive tool to an independent digital hacker capable of identifying vulnerabilities, writing code, and executing attacks. This event marks an unprecedented technological challenge in the cybersecurity landscape, with attack processes now capable of一键式 (one-click) generation from vulnerability discovery to malicious code creation, putting defenders in a more passive position.
OpenAI established a new entity, OpenAI Deployment Company, shifting from technology output to becoming an enterprise service provider, aiming to offer deep engineering support and address technical bottlenecks in AI implementation. At the same time, it acquired the consulting company Tomoro to strengthen its strategic positioning.
OpenAI launched the GPT-5.5-Cyber preview on Thursday, available only to vetted security teams. This cybersecurity-focused version of GPT-5.5 relaxes safety restrictions through targeted training, enabling authorized teams to more efficiently perform vulnerability identification, patch verification, and malware analysis, rather than enhancing offensive capabilities. The release follows Anthropic's launch of Claude My.....
Cybersecurity experts have exposed security risks in the Anthropic company's Claude Desktop application, which silently installs bridging files in seven Chromium-based browsers such as Chrome without user knowledge, and has potential spyware capabilities.