Despite many years of awareness of the problem, the AI and technology field at-large continues to suffer from a severe lack of ethnic, gender, and cultural diversity. This ongoing crisis is particularly pronounced within the technical teams designing the next generation of products – products that impact people from a wide range of backgrounds, often different from the teams developing them.

Part of the problem is underrepresentation. Women and people of color are poorly represented in technical programs at the graduate level, as well as in leadership and management positions at major tech firms. This creates a vicious circle – academic and workplace cultures that many minorities and women find unwelcoming, at the least, loss of potential talent, and a lack of mentors, examples, and advocates for up and coming professionals. 

Moreover, as a recent report by AI Now explains, the diversity problem is not just about academic or professional workstreams. The problem “affects how AI companies work, what products get built, who they are designed to serve, and who benefits from their development.” The report argues that the diversity crisis in the industry and the problems of bias in AI system development and deployment should be understood and addressed as interrelated aspects of the same issue.

To further explore these pervasive challenges and responsibly design AI for the benefit of people and society, DeepMind today gifted £250K to the Partnership on AI (PAI) to support the establishment of a fellowship focused on this task. The fellowship was introduced during a Fireside Chat between Demis Hassabis, DeepMind co-founder & CEO and Terah Lyons, PAI’s founding executive director, at PAI’s All Partner Meeting in London.

The work will begin with consultations across PAI’s Partner organizations and other stakeholders to determine the research topic. Following this, PAI will recruit and host a fellow to complete research focused on understanding and promoting diversity and inclusion in the field of artificial intelligence. The results of the fellow’s work will be published for broad distribution.

The introduction of the fellowship will help the broader AI community make progress against a well-documented problem. We believe that best practices and norms in AI are only possible with the empowered diverse voices of our Partner community. It is essential to design for underrepresented voices if we wish to realize the promise of AI, and we are delighted that DeepMind shares this commitment.

Apply for this fellowship online on our Careers Page.