From Text to Exploits: Vulnerabilities of Apps Using Generative AI and LLMs
summary
Explains how large language models (LLMs) and generative AI are integrated into applications, details common vulnerabilities that arise in such deployments (with concrete examples), and outlines practical developer countermeasures to protect those applications.
AI-written summary · about
chapters (13)
- 0:00LLM introduction
- 4:26How LLM apps work
- 11:11Direct prompt injection
- 14:39Indirect prompt injection
- 16:26Malicious code execution
- 19:33Image-based attacks
- 21:59Custom GPT vulnerabilities
- 24:45Classic defenses limits
- 25:50Advanced defense techniques
- 30:29Dual LLM for security
- 33:40Human confirmation and canaries
- 37:30Best practices summary
- 38:31Conclusion and resources
related talks
46mFrom Text to Flaws: Vulnerabilities in Apps Using Generative AI and LLMs
40mFrom Text to Flaws: Application Vulnerabilities with Generative AI and LLMs
42mFrom Text to Flaws: Application Vulnerabilities with Generative AI and LLMs
51mFrom Text to Flaws: vulnerabilities in applications with Generative AI and LLMs
31mHacking with Words...
49mSecuring AI Apps with the OWASP Top Ten for Large Language Models
45mOverview of Risks, Vulnerabilities and Paths to Safer Use of Generative AI
26mWhen AI Gets Manipulated: Security Risks in LLM Applications