RAII’s Informative Cheat Sheets

The Responsible Artificial Intelligence Institute (RAI) has been working on providing members and the public with a series of cheat sheets that serve as information sheets surrounding various topics such as data quality, fairness definitions, and third-party review, to list a few.

Why Cheat Sheets? In 2020, RAI deployed the Design Assistant, a tool to help members guide and assess the responsible development and deployment of their AI systems. Through user feedback, it became clear that there was a need for supporting context provided to users, allowing one to better understand the material that the Design Assistant assesses. We felt there was a need for a quicker, more digestible way to comprehend important information related to complex topics surrounding the dimensions (Organization Maturity, Accountability, Data, Fairness, Interpretability, Robustness) that the Design Assistant assesses.

While subject information is available online, it tends to be buried in dense text and lends itself to be less accessible. In order to achieve the desired format of information, we landed on using cheat sheets. Here is an example use case using a question from the RAI Design Assistant.A user encounters the question: “Based on the context (domain and tasks) in your system, did you ensure an adequate working definition of “fairness” that you apply in designing AI systems?”A user can then: click on a link to the fairness definitions cheat sheet to see a brief overview of popular fairness definitions used in AI systems, allowing for familiarity with the subject matter to better ground and guide the user’s responses.

A Couple of ExamplesIn order to make information regarding important topics in the responsible AI community easier to access and retain, we opted to structure these cheat sheets as guides which contain items such as checklists, comparison charts, or framed as series of questions posed to the user. The two cheat sheets displayed below are an example of what we are aiming to deliver – clear and brief but informative information on a specific topic within the responsible AI space.

The RAII Independent Review Cheat Sheet, available here: https://portal.responsible.ai/

The RAII Fairness Definitions Cheat Sheet, available here: https://portal.responsible.ai/

The Design Assistant can be used as a way to help prepare members for the RAI Certification program, and these cheat sheets are meant to aid in that preparation by providing an overview of a particular subject in a comprehensible fashion. Moving forward, we hope to expand the bank of cheat sheets we offer to our members and the community. Available cheat sheets can be found in the RAI Community Portal: https://portal.responsible.ai/

Share the Post:

Related Posts

Risks of Generative AI

Responsible AI Institute Webinar Recap (Hosted on April 17, 2024) The proliferation of deepfakes – synthetic media that can manipulate audio, video, and images to...

Responsible AI Institute + AltaML Case Study

Objective    AltaML, an applied artificial intelligence (AI) company that builds powerful AI tools to help businesses optimize performance and mitigate risks, recognized the importance...

Responsible AI Institute Q1 2024 New Members

Multiple Fortune 500 Companies Join RAI Institute; New Member Resources Spotlight Responsible AI Journeys and AI Industry Experts  AUSTIN, TEXAS – Apr. 14, 2024 –...

News, Blog & More

Discover more about our work, stay connected to what’s happening in the responsible AI ecosystem, and read articles written by responsible AI experts. Don’t forget to sign up to our newsletter!