Don’t just map AI risks. Make them understandable.

Introducing RiesgosIA.org

TL;DR

A few weeks ago, I shared that I was starting an AI Safety newsletter. The response was encouraging, and several people offered suggestions, resources, and support. That feedback helped shape something larger than I initially expected.

From newsletter to repository: 

I’ve joined Nicolás Rodríguez at RiesgosIA.org, a project that aims to create a structured, accessible repository of AI risks. We are working together to build a comprehensive resource that helps people explore and understand different categories of AI risks in a clear and organized way.

RiesgosIA is built around a complete taxonomy of AI risks presented as a Periodic Table of Risks, with 118 elements covering topics such as jailbreaks, data poisoning, algorithmic bias, systemic failures, and catastrophic risks. Each element includes a code, description, category, and links to related risks.

What makes RiesgosIA unique The Periodic Table format provides a visual and intuitive way to explore complex AI risks. It helps make abstract concepts more concrete and easier to teach or discuss.

The taxonomy organizes 118 risk elements across technical, societal, and existential categories. Users can explore the interactive table or review the full structured list for deeper analysis.

The platform includes FAQs, learning materials, and resources designed to support people who want to understand AI risks more clearly, whether they are educators, policymakers, developers, or newcomers.

My newsletter will continue as part of this broader effort, offering accessible explanations of AI Safety concepts and contributing to the growing resource library.
 

Why I’m sharing this


The conversations that followed my original post played a direct role in shaping this project. People shared their experiences, pointed out gaps, and suggested collaborations. That input encouraged us to move faster and think more ambitiously.

This is an example of how community feedback can help identify a need and support the people trying to address it.

How you can help If you know communities, researchers, educators, or policymakers who might find this useful, feel free to share it. The periodic table visualization has already proven to be a helpful teaching tool.

If you are working on similar resources or have experience building educational tools for AI Safety, I would be interested in learning from your work.

If you have suggestions, ideas, or constructive criticism that could make the project more useful, we would appreciate it.

Explore the project: riesgosIA.org

Thanks again to everyone who engaged with the first post. This is what grew from it.