From data centers to decision-making technologies, artificial intelligence, or AI, is built at the expense of poor and working-class people. AI injustice occurs when powerful actors use AI, automation, algorithms, and other emerging technologies to entrench and expand economic, racial, and social inequality, often while falsely asserting that these technologies are fairer, more efficient, and innovative. For many, AI is an invisible adversary that denies jobs, cuts healthcare, monitor workspaces, raises rent, and creates insurmountable barriers to a better life. AI exploits natural resources, surveils low-income neighborhoods, and threatens the future of human labor. This bleak reality is not inevitable—it is the result of policymakers ignoring low-income communities attempting to survive AI injustice. As public awareness of AI injustice grows, communities are coming together to demand a different approach to AI governance in which communities lead decision-making on the future of AI and technological innovation in society. We offer this policy framework as a starting point for policymakers and community advocates interested in an approach to governing AI that centers the needs of poor and working-class people. 

This new approach is critical because, to date, policymakers have largely ignored the impact of AI and automation on low-income communities. By some estimates, state lawmakers have introduced over 1,000 bills related to AI and digital policy in the last three years. Unfortunately, very few of these proposals directly address the unique hardships that AI imposes on low-income communities. Instead, policymakers have adopted ineffectual, industry-backed “AI guardrails” that prioritize private profit over public interests. These so-called guardrails typically feature unenforceable rights, inadequate remedies, minimal community input, and virtually no public funding for oversight and enforcement. At best, the guardrails approach grants the public a right to file lawsuits challenging harmful uses of AI. However, even this best-case scenario requires affected individuals to shoulder the financial burden of holding powerful tech companies accountable on their own. This fails to account for the reality that many living in low-income communities have limited access to the resources needed to initiate formal legal proceedings, including money, childcare, transportation, internet access, job flexibility, attorneys, disability accommodations, and language access services. 

“Prevent, Protect, and Repair” is informed by our organization’s experience challenging uses of AI and other algorithmic technologies that disrupt access to basic human needs, such as healthcare, education, and housing. It is also shaped by our conversations with organizations across the country fighting to defend human dignity in the digital age. Most importantly, this framework reflects the experiences of poor and working-class people who confront, survive, and organize against the injustices perpetrated by AI systems that experiment with their lives. “Prevent, Protect, and Repair” aims to translate the aspirations of these voices into a vision of transformative social change.