Summary of The Ai Risk Repository: a Comprehensive Meta-review, Database, and Taxonomy Of Risks From Artificial Intelligence, by Peter Slattery et al.
The AI Risk Repository: A Comprehensive Meta-Review, Database, and Taxonomy of Risks From Artificial Intelligence
by Peter Slattery, Alexander K. Saeri, Emily A. C. Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper, Neil Thompson
First submitted to arxiv on: 14 Aug 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Cryptography and Security (cs.CR); Emerging Technologies (cs.ET); Machine Learning (cs.LG); Systems and Control (eess.SY)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper addresses the lack of shared understanding of Artificial Intelligence (AI) risks by creating an AI Risk Repository, a living database of 777 risks extracted from 43 taxonomies. The repository can be filtered based on two overarching taxonomies and easily accessed, modified, and updated via a website and online spreadsheets. The authors developed the repository through a systematic review of taxonomies and expert consultation, using a best-fit framework synthesis to construct their taxonomies of AI risk. The high-level Causal Taxonomy classifies risks by their causal factors (Entity, Intentionality, and Timing), while the mid-level Domain Taxonomy categorizes risks into seven domains: Discrimination & toxicity, Privacy & security, Misinformation, Malicious actors & misuse, Human-computer interaction, Socioeconomic & environmental, and AI system safety, failures, & limitations. The repository is a comprehensive, extensible, and categorized risk database that can serve as a foundation for a more coordinated approach to defining, auditing, and managing the risks posed by AI systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates an AI Risk Repository to help people understand and manage the risks of Artificial Intelligence (AI). This big book of risks has 777 entries, taken from 43 different lists. It’s like a library where you can find all the information about AI risks in one place. The authors used a special way to combine all these lists into categories that make sense. They made two main groups: Causal Taxonomy and Domain Taxonomy. Causal Taxonomy looks at why something might go wrong, while Domain Taxonomy groups risks together based on what kind of problem they are. For example, there’s one group for problems with AI being mean or unfair to people. There’s another group for when AI systems get broken or don’t work right. This big book is special because it helps us talk about and deal with the risks of AI in a better way. |
