HAI's Commitment to Action
Putting Our Values into Practice
Diversity, equity, accessibility, inclusion, and justice are critical to realizing HAI’s mission to improve the human condition — that is, to develop technologies that support and improve the wellbeing of humanity.
We recognize that anti-Black racism, in particular, has a history in this country that dates back to 1619, is active and present to this day, and must be addressed on many fronts, including within academic institutions. We know universities are microcosms of society and therefore these systemic inequalities are present and persist. HAI acknowledges that confronting these and other inequities are central — and, in fact, are essential — to an AI that is “human-centered.”
HAI recognizes the imperative to align our actions with our values, and take steps toward structural change both in our own institutional culture and our work. We are aware that technologies are not neutral in design or impact, and that there are no machine values, only human values. Because of these truths, bias is baked into many AI systems; but that does not mean that “bias” is simply a problem solved by adjusting data sets or machine learning models. We at HAI understand that there is no silver bullet technological fix for bias, discrimination, and racism. Furthermore, we understand that abstract claims to “fairness” — which presume an equal playing field and equal valuation of human life that unfortunately does not currently exist in society — is an insufficient measure of our efforts not only to “do no harm” but to “do good.” We must acknowledge the fact that academic fields of research and study — including AI — have contributed significantly to the social and economic systems of power and inequity that continue to structure society. Since AI research and its applications have outsized reach and impact, HAI recognizes its special responsibility to challenge those systems.
In June 2020, we strengthened our values statement to more forcefully commit to a more inclusive and just Institute that acknowledges those issues through our leadership, our research, our educational activities, and our outreach. We formed a diversity, inclusion, and equity committee to assess our current efforts; diagnose shortcomings; consult with other departments, student groups, and external organizations; recommend immediate actions and long-term goals; and support the development of best practices including auditing and accountability models and timetables. The committee will continue consulting with faculty and graduate and undergraduate student groups as part of its work. We are also strengthening university and national partnerships in anti-racist practices that we seek to integrate into our work. Moreover, we are committing to an ongoing assessment of our participation in actively promoting equity and fairness at all levels and supporting social justice or in implicitly suppressing these goals through inaction and silence.
We know expressions of solidarity are not enough. Our goals must be actionable and accountable. Our focus here is our commitment to structural change.
If you have questions or would like to learn more, please reach our DEI team
Funding expresses priorities informed by values. Our funded research to date includes a range of work on fairness, equity, and bias in AI technologies.
We commit to:
- Renew support for two postdoctoral fellows per year who work on Ethics & AI, in conjunction with the Center for Ethics in Society.
- Renew funding for the Center for Comparative Studies in Race & Ethnicity’s (CCSRE) Race & Technology Fellows program for another year, as well as commit substantial support for CCSRE’s new Race & Technology Action Lab, which proposes to combine research, teaching, and practice in the humanities and social sciences to advance racial equity in the analysis, production, and deployment of new technologies.
- Fund a new annual HAI visiting artist program specifically focused on artists and creatives who have historically been underrepresented, and work that directly relates AI to issues of diversity, equity, inclusion, civil rights, empathy, and accessibility, or otherwise speaks to the larger questions related to what it means to be “human.”
- Actively encourage, document, and track applications (whether funded or not, and if not, why) for HAI research projects and other scholarly efforts directly related to diversity, equity, and inclusion, in addition to research and projects already funded through our regular mechanisms to those ends.
HAI aims to expand educational opportunities for current and prospective students in the field of AI.
We commit to:
- Enhancing support for AI4ALL, an organization encouraging underrepresented youth in technology paths, by increasing scholarship funds for low-income students attending the Stanford program and creating summer internship opportunities for program alumni, building on the Center for the Study of Language and Information’s (CSLI) Summer Internship Program.
- Hiring a new science education postdoctoral student to conduct research and outreach activities aimed at increasing representation in the technical AI field.
- Supporting the development of the Stanford Computer Science Department’s new embedded ethiCS curriculum, launching in the 2020-21 academic year.
- Collaborating in the promotion and sponsorship of courses and educational initiatives that directly engage with race studies, including theoretical, humantistic, social science, and social justice scholarship as well as historical and political analysis informing students’ understanding of their work in technological fields. Although HAI does not have authority over department curricula, these topics may be well suited for cross-listed courses with other academic units. This includes cross-listing the new English Department undergraduate course AI + ART + ACTIVISM with Symbolic Systems, African & African American Studies, and Comparative Studies in Race & Ethnicity. In addition, these topics could be considered for program recommendations or requirements, analogous to Stanford’s Ethical Reasoning requirement.
- Funding new postdoc positions with the Stanford AI Lab (SAIL) and the Center for the Study of Language and Information (CSLI) that actively seek and encourage applicants of diverse backgrounds.
- Launching a new quarterly lecture series devoted to topics in diversity, equity, and inclusion and co-sponsoring similar talks and events in conjunction with African & African American Studies, the Center for Comparative Studies in Race & Ethnicity, the Program in Science, Technology & Society, the Symbolic Systems Program, and others joining in this effort.
HAI has made policy (engagement with governments and civil society organizations through the means of public policy) a core pillar of the institution to ensure AI is used and governed in a responsible manner. Therefore, HAI has a special obligation to establish strong bonds with organizations that are committed to diversity and equity.
We commit to:
- Proactively engage, listen, and learn from civil society organizations and policymakers such as the Urban League, NAACP, Congressional Black Caucus, and many more.
- Work with diversity and inclusion departments in U.S. federal agencies, as well as state and local governments.
- Ensure diversity in our policy roundtables and workshops.
In STEM fields, participation of underrepresented minorities and especially Black people is still dramatically low in the tech industry and in academia. The common claim that there is an intractable “pipeline” problem is a wholly insufficient response.
We commit to:
- Helping recruit and retain significantly more Black students to the undergraduate and graduate programs associated with HAI; that includes working with and supporting department recruiting efforts at HBCUs and other schools where BIPOC students are matriculating as well as offering application advising support when appropriate.
- Actively transforming an institutional culture that is frequently experienced as inhospitable and sometimes actively hostile to Black students and their research interests by drawing on the best practices and resources mentioned above.
- Exploring ways to collaborate with successful diversity-oriented Stanford college onboarding programs such as Stanford Summer Engineering Academy (SSEA), First-Generation/Low-Income (FLI), and Leland Scholars Program (LSP), as well as promising pilots such as Stanford Summer Session with the Urban Studies Program that can be coordinated with the Haas Center for Public Service and Stanford Pre-Collegiate Studies.
- Helping students stay in the field, which involves engaging the Centers for Equity, Community and Leadership, Vice Provost for Undergraduate Education, Vice Provost for Student Affairs, and the Vice Provost of Graduate Education’s Diversifying Academia Recruiting Excellence (DARE Doctoral Fellowship), Enhancing Diversity in Graduate Education (EDGE Doctoral Fellowship), Stanford Graduate Fellowship and Stanford Interdisciplinary Graduate Fellowship, to enhance support for both undergraduate and graduate students of color, especially Black students. This support will also extend beyond funding to a sense of belonging and professional connection, working with relevant student groups and professional organizations.
- Facilitating structured internship and employment networks and opportunities for alumni and professional mentorships, as well as career placement advising and efforts that go beyond BEAM.
We know that real change requires a rigorous review and oversight of our own internal protocols, practices, and decision-making processes in funding, recruiting, hiring, promotions, fellowships, and everyday exchanges where decisions large and small are made.
Moreover, we know that the distribution of funding alone is an insufficient marker of commitment unless accompanied by an ongoing examination of funding criteria and the demographics and research foci of those we fund.
We are not simply waiting on “good intentions” nor the advocacy of a few to implement these internal reviews; we are building them into our funding protocols and oversight.
We commit to:
- Working closely with the Provost and other departments and programs to support faculty hires whose work focuses on the impact of race in the U.S., including in STEM fields. This, as well as hires of faculty whose research includes the study of race that is not solely about impact, is a priority for HAI and the university.
- Formalizing “equity checks” (e.g., How does it benefit society? Who does it benefit? Who is/might be most impacted? Does it address or advance equity and justice?) in the criteria used to evaluate hiring and funding applications as well as events and activities that HAI supports.
- Creating space and time to listen to and learn from the perspectives and experiences of HAI community members of color — faculty, staff, fellows, students — to create a transparent and inclusive culture within the Institute: HAI leadership regularly participating in anti-bias and social justice work; developing a shared vocabulary around common terms and social concepts; and engaging with historical and social scientific resources to foster deeper understandings of race, racism, and anti-racist work that can also amplify Black voices and other historically marginalized presences. Stanford is a national research hub for these resources through CCSRE, African & African American Studies, and the Digital Civil Society Lab. Additional resources include Raceworks through the Center for Comparative Studies in Race & Ethnicity and other toolkits through Sparqtools from the Social Concepts Lab at the Department of Psychology, Stanford’s Office of Faculty Development, Diversity and Engagement, the School of Medicine, the Graduate School of Business, the Center for Racial Justice at Stanford Law School, the Provost’s IDEAL initiative, and others.
- Creating an online resource library jointly with CCSRE that includes and expands this anti-racist education to include an understanding of intersectionality, recognition of our neurodiversity, and impact of all forms of discrimination including, among others, xenophobia, homophobia, transphobia, and bias based on gender and ability. This understanding includes the relevance of these issues to STEM fields and not just the humanities and social sciences.
- Creating a list to provide university committees charged with naming or renaming spaces and places on campus of the names of Black people, underrepresented minorities, and women who have contributed to the development of AI at Stanford or beyond; this effort is to correct histories of both technology and Silicon Valley that erase their contributions.
- Creating a public, regularly updated HAI Diversity Dashboard based on our internal information as well as data from the Office of Institutional Research and Decision Support and IDEAL dashboards’ aggregated data to help us track grant applications/selections, leadership and staff demographics, number of events related to diversity, etc., to give us evidence-based baselines and metrics for progress.
- Creating a HAI Community Cabinet that includes Black and underrepresented scholars, students, and alumni, as well as active community and external members engaged in issues and topics directly related to AI and issues of diversity, inclusion, accessibility, and community.
- Hiring a HAI Inclusion, Equity, Justice Strategist, a joint position with the Symbolic Systems Program. Crucially, the individual will invite, initiate, and coordinate the commitments we are making today and in the future, keep abreast of current research and resources on campus and nationally, network with campus and community partners, track our progress, provide accountability metrics and audit timetables, and rigorously analyze our funding programs to ensure that relevant projects by people of color, especially Black people, are sought out and supported in both numbers and grant size. The Strategist will structure regular seminars, speaker series, and workshops on the relevant topics, build out diversity-related content and engagement opportunities on the HAI website, ensure feedback loops with the HAI community at large, and generally work with the HAI leadership to ensure diversity efforts become part of our everyday practice and culture.
- Hiring a dedicated staff member in no way should deflect from the fact that diversity cannot be delegated; research is very clear that a high percentage of any group must be actively engaged in anti-bias efforts to be successful. An effective Strategist will ensure that engaging diversity and the openness to critique does not fall to one or a handful of people but is understood as a shared community value and among HAI’s highest priorities.
Our founding committee is comprised of HAI's core leadership. The committee consulted with faculty and student groups on these initial commitments to action and plans to broaden committee membership outside the leadership team. Email HAI-Diversity@stanford.edu to learn more.
William Robertson Coe Professor of Humanities, Department of English, Center for Comparative Studies in Race & Ethnicity, Stanford University
Anand Rajaraman and Venky Harinarayan Professor, School of Engineering, Professor of Computer Science, Stanford University; Vice-Director and Faculty Director of Research, Stanford HAI