AISP Toolkit Feb25 2025 - Flipbook - Page 56
Resources: Data Analysis
Recommendations from the Equitable Data Working Group, 2021, Equitable Data Working
Group
Five Ethical Risks to Consider Before Filling Missing Race and Ethnicity Data, 2021, The
Urban Institute
Community Engagement in Causal Analysis Part 1 & Part 2, 2024, We All Count
CENTERING RACIAL EQUITY THROUGHOUT THE DATA LIFE CYCLE
The SEEDS of Indigenous Population Health Data Linkage, 2021, Rowe, Russo Carroll, Healy,
Rodriguez-Lonebear, & Walker
52
Racial Equity in Use of Algorithms &
Artificial Intelligence
The proliferation of publicly available algorithmic tools, including arti昀椀cial intelligence, is rapidly changing
our relationship with data, generating new ethical questions and considerations for the public sector.
Algorithms are statistical tools that allow for automated problem solving. These tools can put together
data in new ways that have the potential to be leveraged for social good, but they can also intentionally
or inadvertently do harm. There is no such thing as race-neutral algorithms, since this technology
re昀氀ects the biases of those who create them and the data used in their processes. While use of these
tools typically falls under Data Analysis, we treat them as their own stage of the data life cycle because
of their importance in current conversations on ethical data use. We are not experts here, so our goal
is not to give a comprehensive review of these subjects, but rather to put these technologies in the
context of what we do: inform data governance processes that ensure the ethical use of data, particularly
individual-level records and linked administrative data. Below, we provide examples, resources, and a list
of organizations to follow for further ideas and guidance to help you and your organization/community
interrogate these tools with an equity lens.
As always, de昀椀nitions and distinctions are important. For the purposes of this Toolkit, we use
“algorithm” as the umbrella term but also touch on these subtopics: arti昀椀cial intelligence (AI),
machine learning (ML), and deep learning (DL).
