AI-Driven Cyber Threats Escalate: Five Eyes Warn of Open Weight Model Risks

The Five Eyes cybersecurity agencies issued a call-to-action regarding the growing threat posed by AI in cyber attacks. They emphasize that AI is lowering barriers for malicious actors and accelerating the exploitation of vulnerabilities. A recent evaluation showed the Chinese-made Kimi 2.5 model competes with frontier models like Opus 4.5 and GPT-5.2 in structured cybersecurity tasks. Despite not being as capable in vulnerability discovery, these open-weight models are improving rapidly. A report from OALABS Research highlighted a hacker using Opus 4.5 and GPT-5.2 to successfully compromise at least 14 companies. The hacker's operational security was poor, revealing personal information that identified him as a young man from Addis Ababa, Ethiopia. The incident underscores the risks associated with open-weight models in the hands of even less skilled attackers.

Key Points: • Five Eyes agencies warn of AI's role in accelerating cyber threats. • Open-weight models like Kimi 2.5 are closing the gap with frontier models. • A hacker used AI tools to compromise 14 companies despite poor operational security.