From the course: Azure OpenAI Services Security
Join today to access over 24,700 courses taught by industry experts.
OWASP top 10 for LLM applications
From the course: Azure OpenAI Services Security
OWASP top 10 for LLM applications
- [Instructor] The OWASP Top 10 for large language model applications is a project collecting security guidance to help developers, data scientists, and security experts designing and building large language model applications and plugins. Like the other OWASP top 10 lists you might know, the project collects the list of top 10 most critical vulnerabilities for LLM applications from the community. The project includes common examples of each vulnerability, prevention and mitigation strategies, and example attack scenarios. With version 1.0 released in August 23, and version 1.1 in October 23, it's one of the first publicly-available projects that has already produced actionable results in this area. While its primary intention is to support securing AI applications, it can be applied to securing AI usage and AI platforms as well. Prompt Injection occurs when an adversary manipulates the LLM through especially-constructed prompts, causing the model to unknowingly execute arbitrary…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.