Securing AI Integration with API's
summary
Claims that API-AI integrations raise new exploit classes like model extraction and prompt injection, shows a demo of Salt (free dashboard) and Wallarm-style protections spotting incidents, vulnerabilities and shadow APIs, and gives concrete defenses: use UUIDs instead of sequential IDs, enforce TLS and JWT validation with timestamps and refresh tokens, apply least privilege and whitelist responses, and rate limit (example: 100 rps). Also recommends pre-prod scans with dummy data and runtime anomaly detection integrated with Splunk/AppDynamics.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:54New threats in AI age
- 3:18API attack scenarios
- 10:43API security best practices
- 12:28Threat models in AI integration
- 14:59API security building blocks
- 17:26Demo: API security tools
- 19:10Q&A
related talks
24mDiscussion between Mehdi Medjaoui and Troy Leliard,API Security.
18mCommit Secure AI: Embedding Security in Code, Pipelines, and Production
23mTraceable AI
57mAI Applications Security Puzzle
30mNavigating the New AI Security Landscape
25mThe New API Vulnerability Frontier
05mSecuring the Future of APIs: AI-Powered Threat Detection & DevSecOps Integration
39mAPI Security: Acting Now to Enable Security Controls