What are the risks from Artificial Intelligence?

A comprehensive living database of over 700 AI risks categorized by their cause and risk domain.

What is the AI Risk Repository?

The AI Risk Repository has three parts:

  • The AI Risk Database captures 700+ risks extracted from 43 existing frameworks, with quotes and page numbers.
  • The Causal Taxonomy of AI Risks classifies how, when, and why these risks occur.
  • The Domain Taxonomy of AI Risks classifies these risks into seven domains (e.g., “Misinformation”) and 23 subdomains (e.g., “False or misleading information”).

How can I use the Repository?

The AI Risk Repository provides:

  • An accessible overview of the AI risk landscape.
  • A regularly updated source of information about new risks and research.
  • A common frame of reference for researchers, developers, businesses, evaluators, auditors, policymakers, and regulators.
  • A resource to help develop research, curricula, audits, and policy.
  • An easy way to find relevant risks and research.

AI Risk Database

The AI Risk Database links each risk to the source information (paper title, authors), supporting evidence (quotes, page numbers), and to our Causal and Domain Taxonomies. You can copy it on Google Sheets, or OneDrive. Watch our explainer video below.

Search below if you want to explore the risks extracted into our database. This search looks for exact text matches in one field: "Description". It returns information for four fields: "QuickRef", "Risk category", "Risk subcategory", and "Description". For example, try searching for "privacy" to see all risk descriptions which mention this term.

Causal Taxonomy of AI Risks

The Causal Taxonomy of AI risks classifies how, when, and why an AI risk occurs. You can explore the taxonomy (to three levels of depth) in the interactive figure below. Read our preprint for more detail.

Search below if you want to explore how we group risks by cause in our database. This search looks for exact text matches in three fields: "Entity", "Intention" and "Timing". It returns information for seven fields: "QuickRef", "Risk category", "Risk subcategory", "Description", "Entity", "Intent", and "Timing". For instance, try searching for "Pre-deployment" to see all risks of this category.

Domain Taxonomy of AI Risks

The Domain Taxonomy of AI Risks classifies risks from AI into seven domains and 23 subdomains. You can explore the taxonomy (to four levels of depth) in the interactive figure below. Read our preprint for more detail.

Search below if you want to explore how we group risks by domain. This search looks for exact text matches in two fields: "Domain" and "Subdomain". It returns information for six fields: "QuickRef", "Risk category", "Risk subcategory", "Description", "Domain" and "Subdomain". For instance, try searching for "Misinformation" to see all risks categorized in this domain.

How to use the AI Risk Repository

We provide examples of use cases for some key audiences below.

Frequently Asked Questions

Our Team

Acknowledgments

Feedback and useful input: Anka Reuel, Michael Aird, Greg Sadler, Matthjis Maas, Shahar Avin, Taniel Yusef, Elizabeth Cooper, Dane Sherburn, Noemi Dreksler, Uma Kalkar, CSER, GovAI, Nathan Sherburn, Andrew Lucas, Jacinto Estima, Kevin Klyman, Bernd W. Wirtz, Andrew Critch, Lambert Hogenhout, Zhexin Zhang, Ian Eisenberg, Stuart Russel, and Samuel Salzer.

How to engage

Read our preprint, and copy and use our database

Follow MIT FutureTech on social media for updates

AI Risk Repository
© 2024 MIT AI Risk Repository