States Write the Rules, Communities Feel the Impact: What California and Virginia Mean for Black and Brown Futures in the Age of AI
Artificial intelligence is altering daily life faster than federal law can keep up. That leaves states to decide how much protection people receive when automated systems affect their jobs, privacy, safety, and even their identities. California and Virginia have become two of the most active states in this space, but they are writing very different stories about what accountability should look like. For Black and Brown communities, these choices matter. They shape whether AI becomes another tool that reinforces inequality or a system that preserves dignity and civil rights. The intended audience for this analysis includes policymakers, civil rights advocates, and general readers who want to understand how state laws shape everyday experiences.
California’s approach: protecting identity, labor, and youth
California continues to pass some of the most wide‑ranging AI legislation in the country. The Digital Dignity Act would treat using a digital replica to impersonate someone as a form of false impersonation and require platforms to allow people to revoke access to replicas created with generative tools (Barcott, 2026). As of the most recent update, the bill remains active and is moving through the legislative process.
For Black and Brown communities, who already face disproportionate risks of misrepresentation and online harassment, this kind of protection is not abstract. It is a safeguard against deepfake abuse, identity theft, and the weakening of personal agency.
Research on algorithmic bias demonstrates that automated systems often replicate racial disparities present in historical data. Buolamwini and Gebru (2018) found that commercial facial recognition systems had significantly higher error rates when identifying darker‑skinned women compared to lighter‑skinned men. These disparities raise concerns that generative AI tools capable of producing synthetic media may amplify existing patterns of racial misidentification and reputational harm.
California is also focusing on children and adolescents. Several bills address chatbots, youth safety, and the mental health impacts of AI on young people. These proposals acknowledge that Black and Brown youth often enter digital spaces with less support and fewer digital resources than others.
Privacy remains a top concern. A proposed amendment to the California Consumer Privacy Act would require AI operators to provide consumers with access to their personal information, contextual data, and social graph within five business days of a request (Barcott, 2026). For communities long targeted by surveillance and data extraction, this transparency offers a rare glimpse into how their information is used to train and operate AI systems.
The American Civil Liberties Union (2018) found that facial recognition systems were more likely to misidentify people of color, increasing the risk of wrongful stops and arrests. Scholars have also shown that predictive policing systems often rely on historically biased arrest data, reinforcing patterns of over‑policing in Black neighborhoods (Benjamin, 2019; Eubanks, 2018). Greater transparency in data access is therefore directly connected to long‑standing concerns about racialized surveillance.
California is also moving forward on workplace protections. Proposed legislation, such as Senate Bill 947, would prevent employers from relying solely on artificial intelligence to fire or discipline workers. This measure could offer greater transparency and protection for Black and Brown workers, who have historically faced heightened scrutiny and bias in employment settings.
A Brookings Institution report notes that automated decision systems in employment and credit markets can embed structural inequalities unless actively audited for disparate impact (West et al., 2019). Because Black and Brown workers have historically faced discriminatory hiring practices, safeguards against automated termination or discipline are especially significant.
Virginia’s approach: centering responsibility and enforcement
Virginia’s Senate Bill 365 would create the FAIR AI Enforcement Fund, with money distributed by the State Treasurer at the request of the Attorney General (Pekarsky, 2026). This matters for Black and Brown communities because enforcement is often where civil rights protections fail. Without resources, even the strongest laws become symbolic.
If harm comes from an AI system, the bill clarifies that claiming the system acted independently is not a defense (Pekarsky, 2026). This is critical in areas like hiring, housing, and criminal justice, where automated systems have already shown racial bias.
A ProPublica investigation found that one widely used risk assessment tool was more likely to incorrectly label Black defendants as high risk compared to white defendants (Angwin et al., 2016). The broader literature confirms that algorithmic systems can reproduce inequities when accountability mechanisms are weak. Virginia’s refusal to allow an “AI autonomy” defense directly addresses this structural concern.
The bill also calls for clear disclosure of AI models, including updates to training data and terms of service. Importantly, providing disclosure does not protect a company from liability (Pekarsky, 2026). Transparency alone cannot replace responsibility.
Conclusion
As artificial intelligence becomes increasingly embedded in society, states face a choice: will they shape this technology to safeguard justice or allow new inequalities to deepen? California’s protections and Virginia’s emphasis on accountability both point to different ways forward. For Black and Brown communities, these legal choices signal whether AI will deepen inequity or serve as a tool for justice and dignity. A clear next step for other states is to adopt Virginia’s no‑autonomy defense rule and California’s transparency and youth protections, creating a national baseline that ensures AI systems cannot harm communities without accountability.
References
American Civil Liberties Union. (2018). The dawn of robot surveillance: AI, privacy, and civil liberties. https://www.aclu.org
Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias. ProPublica. https://www.propublica.org
Barcott, B. (2026, February 20). AI legislative update: Feb. 20, 2026 – Transparency Coalition. Transparency Coalition.
https://www.transparencycoalition.ai/news/ai-legislative-update-feb20-2026
Benjamin, R. (2019). Race after technology: Abolitionist tools for the new Jim Code. Polity Press.
Buolamwini, J., & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. Proceedings of Machine Learning Research, 81, 1–15.
Eubanks, V. (2018). Automating inequality: How high-tech tools profile, police, and punish the poor. St. Martin’s Press.
Pekarsky, S. G. (2026, February 4). SB365 – 2026 Regular Session. Legislative Information System. https://lis.virginia.gov/bill-details/20261/SB365/text/SB365
West, D. M., Whittaker, M., & Crawford, K. (2019). Discriminating systems: Gender, race, and power in AI. Brookings Institution. https://www.brookings.edu