Sourcebow·6d agoAI Agent Guardrails for Production The Cost of Giving AI Too Much AccessAn AI agent instructed to issue an $1,800 refund could turn a malicious prompt into a financial loss if the system allows it to execute the…A response icon1A response icon1
Sujal Thapa·Oct 24How to keep AWS Accounts safe from AI Attacks, a SME perspectiveWith AI Cybersecurity risks shaking the digital world recently, those who manage cloud infrastructure can’t help but feel a bit uneasy…
InAI SecuritybySanjeev Jaiswal (Jassi)·Oct 19AI Safety vs AI Security that will clear the fight of two teams!The Definitive Reference Guide to the Distinction, Overlaps, Team Responsibilities, and Governance between AI security and AI Safety team
Imogen Hooton (Armstrong)·Oct 14What You Should Be Reading to Change How You Think About AI ThreatsNo speculation, no fearmongering, and yet… the reality is scarier than you think.
Nuno Roberto·Oct 6I Mapped AI Cyber Risk to a Process Model Older Than Most of My Engineers (2/5)Part two of five on AI security across the lifecycle, covering why CRISP-DM is the right spine and what security does before a single model…
Rob Enderle·Sep 17Taming the Beast: Inside Microsoft’s Desperate Bid to Control AIEvery so often, the tech industry realizes it has built something it doesn’t entirely understand — and cannot reliably control. We saw it…A response icon1A response icon1
Tom Vogel·Sep 15AI in Your Product: A Token Guzzler and a Security RiskDo you really need a chatbot in your software, or is an FAQ page good enough? To answer this question, consider token costs and security…
Manish Prasad·Sep 9OpenAI’s Agents Are Posting on Random Wikis to Cheat on TestsFor six weeks this spring, a swarm of autonomous OpenAI agents took over a 25-year-old German programming wiki. They posted roughly 18,000…
