It’s an exciting time in the world of responsible AI. Just last week, the RAI Institute team traveled to London to participate in London Tech Week and the CogX Festival, where we announced RAI Institute’s expansion into the UK and EU and shared the news of our accreditation pilot launch with SCC with wider audiences. At the international level, we’re also seeing major developments in AI-related regulation, such as Canada’s draft Artificial Intelligence and Data Act (AIDA).

As part of this momentum, we’re excited to share three just-released documents with our global community: the RAI Institute Whitepaper, Certification Guidebook, and Certification Program Scheme Document Sample. Together, these documents provide insight into RAI Institute’s approach and work thus far at the forefront of responsible AI certification. You can access the documents on our website here.

We’re thrilled to share these resources with you, our RAI Institute community, to illustrate who we are and what we do in more detail. As you’re checking out our new materials, we recommend looking at our whitepaper first as it offers an overview of our work from end-to-end. Then, for more information, read our guidebook and scheme sample, which flesh out our certification program in detail.

Why are these three documents important?

The RAI Institute Whitepaper provides an overview of our Certification Program. It also describes the Certification Program’s alignment with AI laws, regulations, principles, research, and practitioner insights, RAI Institute’s initial focus industries and functions, and the involvement of the RAI Institute community.

Why it matters: Our whitepaper is a critical resource for anyone interested in learning more about RAI Institute’s journey and work, including our foundation, our approach, and what our assessment tools look like in practice.

The RAI Institute Certification Guidebook provides background and requirements for the Certification Program, RAI Institute’s processes for development, amendment, and governance, and organization- and auditor-specific information related to certification and recertification.

Why it matters: Our guidebook provides a detailed overview of our process and principles to provide transparency to our stakeholders and demonstrate the wide body of research upon which our work is built.

The RAI Institute Certification Program Scheme Document Sample is a selection of eight questions from the broader RAI Institute Certification Assessment Program Scheme Document, which contains the 89 scored assessment questions of the Certification Assessment. This sample document features assessment questions from each of RAI Institute’s Implementation Dimensions.

Why it matters: Our scheme document represents a breakdown of the well-researched intent and rationale behind every question we ask in our certification assessment. We have released this sample to give the community a tangible preview of our work as we continue to finetune it.

What’s next

The development and release of these three materials mark another huge milestone in our journey to widely implementing our first-of-its-kind RAI Institute certification.

Our Certification Program and conformity assessment scheme is based on our Implementation Framework, calibrated to specific use cases, and aligned with principles, laws, guidelines, research, and best practices relevant to the industry, function, and jurisdiction in which an AI system is developed and deployed.

We have already been successful in supporting the responsible development and procurement of AI systems in use cases related to lending and collections, human resources, health diagnosis and treatment, and access to health care. Our initial work has focused on the US, UK, and Canada, and we anticipate that it will encompass other geographies, too.

Building on these three documents, we will continue to calibrate our assessment to real-world uses and the latest developments in the field. We’re also excited to move forward this year with our Artificial Intelligence Certification Pilot with the Standards Council of Canada (SCC), which you can learn more about here.

As we continue to evolve our work and strive toward the adoption of better, safer, and fairer AI, we are grateful for the insights and support of the RAI Institute community. Thanks for joining us on this journey.