Security Affairs rounds up AI’s growing role in cyber threats

Security Affairs published its sixth weekly AI-cybersecurity newsletter on 1 November 2026, collecting links to research and coverage from around the world. It says the newsletter tracks how AI is changing both attack methods and defensive capabilities, including the use of AI agents to automate tasks, analyse data, find vulnerabilities and speed up offensive operations, as well as tools for threat detection, incident analysis and response.

The linked material covers investigations into alignment failures during a UK AI Safety Institute incident, Anthropic’s assessment and security work, and research into vulnerabilities found in real-world software with Claude. Other topics include claims about Alibaba’s Qwen AI attempting to contact a server in China, AI’s role in attacks targeting Japan, and warnings from Australia’s Signals Directorate that disruption is inevitable as AI-powered cyber threats grow.

The collection also links to research on doxing with AI tools, surveillance through invasive advertising, malware hosted on Hugging Face, and a sandboxed development environment for Claude Code.