- The number of initiatives advancing equity and inclusion has increased substantially over the past decades.
- Technology companies have an added complexity for equity and inclusion challenges unique to the design, development, deployment and use of technologies such as artificial intelligence (AI).
- This report paints a comprehensive picture of challenges and opportunities for improvements in equity and inclusion across the AI development life cycle and governance ecosystem.
- Read the framework here
San Francisco, USA, 29 June 2022 - The World Economic Forum published "A Blueprint for Equity and Inclusion in AI" to identify challenges and solutions for equity and inclusion across each phase of the AI technical life cycle and its governance ecosystem. The report proposes best practices for each step within AI development, but also gives a holistic approach that connects each stage to benefit the broader ecosystem.
"Bias in AI is a real concern," said Kay Firth-Butterfield, Head of the AI and Machine Learning Platform at the World Economic Forum. "Over the course of 10 months, the diverse group that makes up the Global Future Council on Artificial Intelligence for Humanity mapped the vast number of bias-related risks in the AI lifecycle and developed a blueprint that sets the foundation for what it means to have an inclusive AI ecosystem. Building on AI ethics, this blueprint connects the dots where they previously have not been connected."
The blueprint is the result of a collaborative effort of members of the World Economic Forum Global Future Council (GFC) on Artificial Intelligence for Humanity and includes input from interviews held with people of diverse backgrounds and representatives of institutions in the AI diversity, equity and inclusion ecosystem. "This unique report, created by a global community of leaders, provides a lens into a more inclusive AI ecosystem. Collectively, we created a roadmap with actionable items to develop more equitable AI impacting all communities," said Susan Gonzales, Founder & CEO, AIandYou.
The lack of inclusive AI has proven harmful to businesses (leading to financial or reputational harms) and certain communities, in particular ethnic, cultural, linguistic, ideological, and generational minorities. Therefore, the blueprint underlines the importance of adopting equity and inclusion strategies for AI through its pre-integrated AI ethics approach. "For Māori, including different skills, backgrounds, thinking and values at any decision-making table is the only way to harness power for a sustainable, collective 'good', to better ensure a representative, responsible world, let alone an AI lifecycle. That is the evolution, not a revolution that is needed," said Sara Stratton, Founder, Māori Lab
The blueprint aims to provide a better understanding of how equity and inclusion, by design, can be adopted at two levels: the ecosystem (broad) and the AI lifecycle (stages of technical development). "Inclusion is too often an afterthought in how we approach the visioning, design and implementation of new digital capabilities. The result? The work being driven by a small, privileged group and those without a stake in the process being left out of the benefits. This blueprint provides ways to counteract this trend in the domain of AI, where if acted upon could lead to an acceleration toward a more inclusive society at large," said Nadjia Yousif, Managing Director and Partner, Boston Consulting Group.
The intended audience of the framework includes managers and teams responsible for the different stages of AI development, as well as decisionmakers from both the private and public sectors part of the AI ecosystem. "It is no longer enough to invite underrepresented communities to a table and label it inclusion, this blueprint demonstrates what is truly required to honour the word 'inclusion' and its significant meaning through how organisations can evolve community engagement from tick-boxing to transformative," said Julia Arnott-Neenee, Co-Founder and Director, PeopleForPeople_NZ.
The Blueprint for Equity and Inclusion in AI was developed by the GFC through monthly meetings of the council members, anonymous stakeholder interviews, and a workshop held with members of the World Economic Forum's Global AI Action Alliance.
