Creating technologies that are equitable and just is a crucial aspect of the ongoing evolution in artificial intelligence (AI). As AI systems become more integrated into our institutions, from healthcare and education to employment, the imperative to ensure these systems are designed and implemented in ways that promote racial justice and equity has never been more pronounced. Today, racial bias can be seen in applications of AI in employment, the law, and search behaviors. For example, AI models that recommend jobs and educational opportunities on job boards and social media platforms have been found to recommend roles to BIPOC professionals that are below their current qualifications (https://www.thomsonreuters.com/en-us/posts/legal/ai-enabled-anti-black-bias/), while applications from candidates with Black names are often rejected by HR hiring systems (https://www.newamerica.org/oti/blog/ai-discrimination-in-hiring-and-what-we-can-do-about-it/). While the law is catching up with AI tools and banning AI-driven forms of discrimination, we know that regulations are not a sufficient tool for addressing systemic, structural biases.
As AI continues to be trained on data influenced by racial disparities, we face a turning point where data-driven decision-making by AI risks reproducing the inequities we see today, at scale.
Thankfully, several AI organizations and programs are at the forefront of integrating racial justice into their core mission, demonstrating how technology can be harnessed to foster a more equitable society. We’re here to celebrate these important initiatives and call for more attention to issues of racial equity in AI.
Data for Black Lives
Description: Data for Black Lives is a movement of activists, organizers, and scientists committed to the mission of using data science to create concrete and measurable change in the lives of Black people. The initiative seeks to use data and AI to fight bias, build progressive movements, and promote racial equity. Their work includes research, policy advocacy, and the development of tools that empower communities to challenge systemic racism and inequality.
Algorithmic Justice League
Description: Founded by Joy Buolamwini, the Algorithmic Justice League (AJL) is an organization that aims to raise awareness about the social implications of AI, focusing on how these technologies can perpetuate racial bias and discrimination. AJL conducts research, advocacy, and art to promote more equitable and accountable AI. Their initiatives include auditing commercial AI systems for bias and working with policymakers to ensure the development of more just and equitable AI technologies.
AI for Good's Workforce, Diversity & Transparent Workplaces with AI
Description: AI for Good is an organization that leverages AI technologies to tackle some of the world's most pressing challenges, including racial injustice. Their Racial Equity and Justice Initiative focuses on developing AI solutions that address systemic racism and inequality in the workplace. Racism is one of the issues they address.
The Greenlining Institute’s Technology Equity Program
The Greenlining Institute works to bring investment and growth to communities of color through a range of initiatives. One of these endeavors is the Technology Equity Program. Including AI but also addressing infrastructure, access, and career opportunities, the Technology Equity Program aims to increase access to capital and economic growth across the board. They are working to address algorithmic redlining through advocacy, education, and research.
Latimer.AI
Latimer is a competitor of ChatGPT and other large language models. Unlike the typical LLM, Latimer is trained on more diverse data. This is important, because large language models, like other machine learning and AI technologies, depending on data from the real world to “learn” basic facts and skills. If the data used to train an LLM is limited to data relating to people of privileged identities, or biased by a privileged or narrow lens, it will reproduce existing societal biases. To give an example, an AI trained on data relating to crimes in Black neighborhoods, but not on arts, culture, and business leadership in the same neighborhoods, will produce new content that talks about crime, not culture or business leadership. That is why diverse datasets are so important in AI. Latimer addresses this need by training its LLM with a wide range of representative texts, to present a more accurate, fuller, and more authoritative picture of communities of color.
These initiatives represent just a fraction of the work being done at the intersection of AI and racial justice. By focusing on the development of equitable and just AI systems, these projects are paving the way for a future where technology serves as a force for good, helping to dismantle systemic racism and promote a more equitable society. As AI continues to evolve, the commitment of these initiatives to racial justice serves as a critical reminder of the importance algorithmic justice, as well as providing a clear path forward to a more just application of AI across sectors.