AI Threat
How attackers use AI-generated video and cloned voices to impersonate executives and authorise fraudulent transfers; the Arup ~US$25.6M case.
AI Threat
How LLMs write clean, localised, personalised phishing and BEC lures at scale, erasing the spelling and grammar tells defenders relied on.
AI Threat
How direct and indirect prompt injection makes LLM copilots and agents follow attacker instructions hidden in data; OWASP LLM01, ATLAS AML.T0051.
AI Threat
The criminal 'dark LLM' market (WormGPT, FraudGPT) and documented nation-state abuse of mainstream models; what is real versus hype.
AI Threat
How adversaries actually use AI in 2024-2026, what is hype versus real, and the frameworks (ATLAS, OWASP, NIST, CISA/NCSC) that structure the defence.