Posts on ai security

Fintechs use AI for fraud detection, alert triage and automation and attackers use it against them. What the FCA expects and the controls that actually reduce risk.

AI security protects the AI systems you run and defends against attackers using AI against you. Risks, controls, and the UK standards that apply in 2026.

How CyberSecEval 2 scores LLM security risk, GPT-4, Llama 3, the arXiv 2404.13161 benchmarks explained, and what the results mean for your AI stack.