- T10 FOR GEN AI
- Project Initiatives
Initiatives
The goal of initiatives within the project are to address specific areas, of education and research to create practical, executable resources and insights in support of the overall project goals through focused working groups. Each inititive charter is reviewed and approved as outlined in the OWASP Top 10 for LLM Project governance.
AI Cyber Threat Intelligence
Limited actionable data exists in understanding how different LLMS are being leveraged in exploit generation. This initiative aims to explore the capabilities and risks associated with generating day-one vulnerabilities’ exploits using various Large Language Models (LLMs), including those lacking ethical guardrails.
Guidance & Resources
Get Involved
- Slack: #team-llm_AI-cti
- Initiative Charter
Initiative Lead(s)
Working Group
Secure AI Adoption
The Secure AI Adoption Initiative forms a Center of Excellence (CoE) to enhance security frameworks, governance policies, and cross-departmental collaboration for Large Language Models (LLMs) and generative AI. Through strategic planning, training, and the development of standardized protocols, the initiative ensures that AI applications are adopted safely, ethically, and securely within organizations.
Guidance & Resources
Get Involved
Initiative Lead(s)
- Scott Clinton
Working Group
AI Red Teaming & Evaluation
This project establishes comprehensive AI Red Teaming and evaluation guidelines for Large Language Models (LLMs), addressing security vulnerabilities, bias, and user trust. By collaborating with partners and leveraging real-world testing, the initiative will provide a standardized methodology for AI Red Teaming, including benchmarks, tools, and frameworks to boost cybersecurity defenses.
Guidance & Resources
Get Involved
- Slack: team-llm-redteam
- Weekly Meeting - Thurs 9am PST
- Initiative Charter
Initiative Lead(s)
Working Group
Risk and Exploit Data Gathering, Mapping
This initiative gathers real-world data on vulnerabilities and risks associated with Large Language Models (LLMs), supporting the update of the OWASP Top 10 for LLMs. In attition this initives maintains mappings between the Top 10 for LLM and other security frameworks. Through a robust data collection methodology, the initiative seeks to enhance AI security guidelines and provide valuable insights for organizations to strengthen their LLM-based systems.
Guidance & Resources
Get Involved
Initiative Lead(s)
Working Group
Agentic Security Initiative
The Agentic Security Research Initiative explores the emerging security implications of agentic systems, particularly those utilizing advanced frameworks (e.g., LangGraph, AutoGPT, CrewAI) and novel capabilities like Llama 3’s agentic features.
Guidance & Resources
Get Involved
Initiative Lead(s)
Working Group
Resources
Agentic AI – Threats and Mitigations
Agentic AI represents an advancement in autonomous systems, increasingly enabled by large language models (LLMs) and generative AI. While agentic AI predates modern LLMs,...
LLM and Gen AI Data Security Best Practices
The rapid proliferation of Large Language Models (LLMs) across various industries has highlighted the critical need for advanced data security practices. As these AI...
GenAI Red Teaming Guide
This guide outlines the critical components of GenAI Red Teaming, with actionable insights for cybersecurity professionals, AI/ML engineers, Red Team practitioners, risk managers, adversarial...
Initiative Blogs
Announcing the OWASP LLM and Gen AI Security Project Initiative for Securing Agentic Applications
The OWASP Foundation is thrilled to announce the launch of the Agentic Security Initiative from the LLM and Generative AI Security Project to tackle...
Research Initiative: AI Red Teaming & Evaluation
Red Teaming: The Power of Adversarial Thinking in AI Security (AI hackers, tech wizards, and code sorcerers, we need you!) This is your invitation...
Research Initiative – Securing and Scrutinizing LLMS in Exploit Generation
Challenge Currently limited actionable data exists in understanding how different LLMS are being leveraged in exploit generation, and what mechanisms can be used to...
