About the Project
About The LLM Risk Atlas
The LLM Risk Atlas is a practical educational project about the ways large language models and AI systems can fail, mislead, or create risk when their outputs are trusted too quickly.
The goal is not to be anti-AI. The goal is to help people use AI more responsibly by understanding common failure modes, recognizing warning signs, and applying practical safeguards before acting on AI output.
Why this project exists
AI tools can be useful for learning, writing, coding, analysis, planning, and security research. But they can also produce confident answers that are unsupported, outdated, incomplete, or wrong. This project turns those risks into plain-English explanations, examples, checklists, and role-specific guidance.
Who it is for
The Atlas is designed for students, developers, cybersecurity analysts, business leaders, compliance and risk teams, and general AI users who want to understand where AI output can go wrong and how to verify it more carefully.
Project perspective
The project approaches AI risk from a practical cybersecurity, governance, and technical education perspective. It emphasizes source verification, privacy awareness, human review, least privilege, safe tool use, and clear communication.
Creator
Created by Leni Halaapiapi as a professional portfolio project combining AI literacy, cybersecurity risk thinking, technical writing, frontend development, responsible AI education, and governance, risk, and compliance awareness.
Design principle
Practical, not alarmist.
The LLM Risk Atlas treats AI as a powerful tool that still requires human judgment. The site focuses on practical verification habits, realistic examples, and clear mitigations instead of hype, fearmongering, or vendor-specific marketing.