On June 23, 2020, the Harvard Data Science Initiative released an Anti-Racism Statement in response to the murders of George Floyd, Breonna Taylor, Ahmaud Arbery and the outpouring of public anger their murders impelled. In that Statement, we committed to “identify concrete ways in which we, as a Harvard data science community, can support research programs and education that expose bias in the way data and data science is used, as well as advance research into fair and responsible data science. We also want to take action towards promoting diversity across our governance, research, and education programs.” We write to you today with the initial outcome of our efforts and an invitation to a community Town Hall meeting on Thursday, September 10 to learn more.
In the academic year 2020/2021, the Harvard Data Science Initiative will:
Establish a new program around data science and bias: Bias2
The HDSI’s new Bias2 Program will support research, feature speakers, and engage the data science community towards using data science to uncover bias, and understanding and combating the use of badly-conceived data science that can reinforce bias and inequity. As part of this program, the HDSI will:
Launch the Bias2 Seminar Series, monthly webinars featuring speakers who shed light on different types of bias through the use of data science, as well as speakers who raise bias concerns around problems with data, algorithms, and data-science methodology. We will also feature speakers whose work more generally focuses on public interest technology, including issues and concerns that are broadly related to racism and technology. Full details are online at https://datascience.harvard.edu/bias.
Activate $250,000 to support faculty-led research that fits within the agenda of using data science to uncover bias and addressing bias in data science. More details will be forthcoming at https://datascience.harvard.edu/bias.
Expand HDSI’s summer Public Service Data Science Graduate Fellowships to support a total of three fellows in 2021, at least one of whom will be engaged in work that combats systemic racism.
Publish a special issue of the Harvard Data Science Review on Responsible Data Science
The Harvard Data Science Review will establish a special collection of articles that explore and promote responsible data science. Featuring contributions by expert leaders representing academia, government, and industry, and addressing algorithmic fairness, social justice, and social welfare, this special collection will be established during the coming academic year and will encourage the entire data science community to make an ongoing commitment to addressing these important societal issues.
Revise our HDSI Postdoctoral Fellow recruitment practices
HDSI Postdoctoral Fellows are outstanding early career researchers whose work spans Harvard’s schools and disciplines. They represent the future of data science and their voices must include diverse perspectives and experience. We will revise our recruitment practices to increase outreach to applicants traditionally under-represented in the field and to provide a more welcoming environment to scholars of all races, genders, ethnic and socio-economic backgrounds through mechanisms such as flexibility around appointment timing and funding to help offset relocation costs.
Improve transparency around HDSI programs and governance, and provide resources to support diversity and inclusion
We bear a responsibility to combat bias and racism not just in our programmatic activities, but also in our governance, administration, and community. Over the coming weeks, we will publish on our website information about how we review applications for funding, including postdoctoral fellowships. We will review our governance structures to identify areas where we can include and amplify diverse voices. And, we will provide information about resources to support engagement and inclusion.