Prioritize AI security controls by sorting scenarios into likelihood and impact quadrants.
Sort each AI security scenario by likelihood and impact in a typical enterprise environment. High likelihood + High impact High likelihood + Lower impact Lower likelihood + High impact Lower likelihood + Lower impact Public support bot receives prompt-injection probes every week and can reveal ticket metadata Internal assistant occasionally gives outdated CVE summaries but cannot write to systems Security agent can update firewall rules if a rare prompt/tool chain succeeds Prototype chatbot for training uses synthetic data and has no external users Analysts frequently paste real logs into AI and sometimes include customer identifiers Model response formatting sometimes breaks…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in