Why securing models without securing APIs creates a false sense of control
summary
India ranks highest for reported API incidents and APIs tied to AI represent 43% of attacks, while 23% of visible APIs still hide whether they carry sensitive PII. Unauthenticated error responses and leaked GraphQL schemas expose data without breaking auth, and query-strings leak bank numbers. Gartner expects AI to drive >30% of API traffic growth, magnifying existing gaps. Practical controls are: build an API inventory and classify sensitive endpoints, shift checks into CI/CD, and add runtime protection to catch the remaining 1%. Mythos proof-of-concept shows attacks go from discovery to exploit in minutes.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:11APIs in modern apps
- 3:17API security trends in APJ
- 6:03Real-world API failures
- 9:33Agents amplify API risks
- 10:57Core API security pillars
- 15:30Three API security questions
- 16:33Frontier LLMs and API attacks
- 19:25Q&A
related talks
24mDiscussion between Mehdi Medjaoui and Troy Leliard,API Security.
26mAPI and Application Security - The Future
24mAPI security, mitigate API abuses using machine learning
19mAvoid These API Security Pitfalls: Expert Insights
23mTraceable AI
25mThe New API Vulnerability Frontier
58mA Risk based Approach to API Security
01mOn The Value Of API Security Focus