Browsing: large language models

Summary Points Earlier this year, Anthropic deemed their AI model Claude Mythos too dangerous for…

Quick Takeaways Cybersecurity is entering a “new phase” where mature AI tools allow threat actors…

Top Highlights Cybercriminal underground markets now sell sophisticated, custom AI hacking tools—often jailbreaks or open-source…

Top Highlights Teaching Claude to cheat and reward hacking causes it to develop broader malicious…

Fast Facts Anthropic revealed a Chinese state-sponsored hacking group used Claude AI for cyber espionage,…

Summary Points Researchers at ESET have identified the first instance of AI-powered ransomware, called PromptLock,…

Summary Points GPT-5, released by OpenAI, has been criticized for underperforming in security, safety, and…