Artificial intelligence is often framed as a frontier technology, but its most consequential effects unfold in everyday life. Governance failures, ranging from weak oversight in U.S. public‑sector systems to the opacity of large‑scale biometric programs and the expansive surveillance practices common in authoritarian and hybrid regimes, show how AI can deepen existing inequities when deployed without safeguards. When designed and governed with equity at the center, however, AI can broaden access to education, employment, and essential public services. This brief examines how these contrasting governance choices shape outcomes across different political contexts.  

This cause‑and‑effect relationship is especially visible for Black communities in the United  States. A national study by Jobs for the Future (JFF) found that 83 percent of Black respondents were familiar with AI, and 56 percent used AI weekly for learning, work, or creative projects  (Swartsel, 2025). High engagement should lead to expanded opportunity. Instead, inadequate governance often leads to biased outputs, misrepresentation, and unequal burdens. 
 

At the same time, global developments—particularly in multilingual, resource‑constrained environments—demonstrate that equitable outcomes are possible when governance is intentional. India’s digital public infrastructure (DPI) shows how large‑scale systems can expand access, while also revealing the risks of exclusion when rights protections are weak (Elias,  2026).  
 

Comparative politics research reinforces this point: AI is not merely a technological field but a domain of political power shaped by institutions, ideologies, and international hierarchies  (Freedom House, 2024). When governance is strong, AI can advance justice. When governance is weak or authoritarian, AI becomes a tool of surveillance, exclusion, and control. 
 

Cause: Biased Data and Weak Oversight  

Effect: Harmful Outcomes for Black Learners, Workers, and Journalists 

Black Learners Are Early Adopters—But Governance Has Not Kept Pace  

Black learners and workers are among the most active users of AI tools (Swartsel, 2025). This should produce positive outcomes: improved study support, expanded career pathways, and greater creative capacity. Instead, weak governance allows biased systems to shape their experiences.  

 

Effect 1: Reinforcement of Criminal Justice Bias  

A composite scenario, based on patterns identified in JFF’s findings, illustrates this dynamic. In this hypothetical example, a Black criminal justice student uses AI to study case law and policing scenarios. Because the system is trained on biased criminal justice data, it repeatedly generates examples centered on Black neighborhoods—even when no demographic information is provided.

Cause: AI systems trained on historical arrest data.  

Effect: Reinforcement of predictive policing patterns that disproportionately target Black communities.  

Instead of helping him challenge inequities as a future public defender, the AI reproduced the very assumptions he sought to dismantle.
 

Effect 2: Misrepresentation in Journalism  

Composite Scenario: A hypothetical example illustrates how these issues appear in newsroom settings. In this scenario, a Black journalist relies on AI transcription tools to manage heavy workloads. The tools frequently misinterpret African American English and regional dialects, requiring hours of correction. AI‑generated summaries flatten cultural nuance, stripping essential context from stories rooted in Black experiences.

Cause: AI systems are trained on limited linguistic diversity.  

Effect: Higher error rates for journalists of color and distorted representation of their communities (Radsch & Lee, 2024).
 

These outcomes are not accidental—they are the predictable effect of governance that fails to require diversified training data, auditing, and accountability.  

The NAACP (2025) has warned that without strong protections, AI will continue to misrepresent  Black voices, reinforce discrimination, and undermine political power.  

Cause: Political Regime Type  

Effect: Divergent AI Governance Models
 

Comparative politics research shows that political systems shape AI governance in predictable ways (Freedom House, 2024).  

Democratic Regimes: Rights‑Based Governance Produces Accountability  

Democracies such as the European Union emphasize transparency, data protection, and regulatory oversight. These systems treat AI as a potential disruptor that must be constrained through legal safeguards.  

Cause: Strong institutions and rights‑based political culture.  

Effect: AI governance that prioritizes civil liberties and public accountability.  

 

The United States has historically lacked a unified AI governance framework, but President Biden’s 2023 Executive Order marked a significant shift. The order reframed AI as a civil rights and national security issue, directing agencies to adopt the NIST AI Risk Management Framework and develop guidelines addressing privacy, workplace monitoring, and algorithmic discrimination (Kalnina, 2025). With the change in administration, however, several EO‑driven initiatives have been slowed, revised, or rolled back, creating renewed uncertainty about the direction and coherence of federal AI governance.

This shift demonstrates how governance choices can directly shape AI’s impact on marginalized communities.
 

Authoritarian Regimes: AI as a Tool of Control  

Authoritarian states such as China and Rwanda deploy AI to strengthen surveillance and political control. In China, AI is integrated into the Social Credit System, public‑security platforms, and expansive facial‑recognition networks used for population monitoring and protest suppression. Rwanda, while often categorized as a competitive‑authoritarian system, exhibits governance features that align with authoritarian practice: concentrated executive power, restrictions on independent media, tight control over civil society, and a growing digital‑surveillance infrastructure used to monitor political opponents. These characteristics justify its inclusion alongside more consolidated authoritarian regimes in this analysis. Nigeria, by contrast, reflects a hybrid regime context in which AI tools are increasingly used for policing and identity verification, but with weaker institutional safeguards and inconsistent oversight. Across these cases, AI systems function as extensions of state surveillance and administrative control, though the scale and sophistication vary by regime type.

Cause: Centralized power and limited civil liberties.  

Effect: AI systems that reinforce state authority rather than protect citizens.
 

Hybrid Regimes: Modernization Without Legitimacy  

Hybrid regimes like Nigeria pursue technological modernization but lack the institutional capacity to regulate AI effectively.  

Cause: Weak oversight, corruption, and contested legitimacy.  

Effect: AI systems that magnify existing inequalities and governance failures.  

Across all regime types, AI serves as a mirror of political values and a motor accelerating their effects.  

Cause: Infrastructure Without Safeguards  

Effect: Uneven Access and New Forms of Exclusion
 

India’s digital public infrastructure (DPI) offers a powerful example of how large‑scale systems can expand access. Platforms such as Aadhaar, UPI, and DigiLocker have enabled millions of people to obtain identification, access financial services, and manage digital records (Elias,  2026). Yet exclusion persists: early Aadhaar audits found biometric authentication failure rates of 6–8 percent in some states, disproportionately affecting older adults, manual laborers, and low‑income residents. These gaps show how access‑expanding systems can still reproduce inequities when safeguards are weak.  

Cause: Interoperable, open digital infrastructure.  

Effect: Expanded access to essential services.
 

However, India’s experience also reveals the consequences of weak rights protections. Cause: Uneven documentation, privacy gaps, and limited grievance mechanisms. Effect: Exclusion of marginalized communities and vulnerability to misuse.  

For minority communities in the United States, India’s model demonstrates that access alone is insufficient. Infrastructure must be paired with strong rights protections to prevent inequitable outcomes.  

Cause: AI Adoption in Education and Employment  

Effect: Expanded Opportunity—When Governance Is Strong
 

AI tools are increasingly integrated into higher education for tutoring, advising, admissions, and administrative support. 78% of college administrators report positive impacts on teaching and operations (Hargrove, 2024).  

Cause: AI‑enabled academic support systems.  

Effect: Reduced administrative burdens and expanded student support—especially at minority‑serving institutions.  

AI is also reshaping recruitment. Machine learning systems can analyze large applicant pools,  identify qualified candidates, and reduce human bias. IBM reports that AI‑enabled recruitment tools can improve diversity by focusing on skills rather than demographic characteristics (Finn &  Downie, 2025).  

Cause: Skills‑based AI hiring tools.  

Effect: Expanded access to higher‑skilled employment pathways for minority workers. 
 

These positive outcomes demonstrate that equitable AI is possible—but only when governance is intentional.  

Cause: Culturally Responsive Design  

Effect: More Equitable AI Systems
 

Community‑centered AI models such as ChatBlackGPT show how culturally responsive design can produce more equitable outcomes. Compared to ChatGPT, ChatBlackGPT provided more culturally relevant resources, explicitly acknowledged identity, and recommended support for  Black‑owned businesses across all user interactions (Egede, 2024). Achieving equity in AI  governance requires this kind of intentional, community‑rooted design.  

Cause: AI systems are designed with cultural context and identity in mind. 

Effect: More accurate, empathetic, and equitable user experiences.  

This model demonstrates that equitable AI requires not only governance but also design choices that reflect lived experiences.
 

Conclusion: Equity Is the Effect of Intentional  Governance  

Artificial intelligence does not inherently produce inequity; inequity is the effect of governance failures. When AI systems are trained on biased data, deployed without oversight, or designed without cultural context, they reproduce and amplify structural inequities. When governance is strong, through rights‑based frameworks, diversified data, and culturally responsive design, AI can expand opportunity.
 

Equity in AI governance must be earned through intentional design, inclusive policymaking, and persistent oversight. In addition to broad principles, this requires concrete policy action. Federal and state agencies could adopt a mandatory algorithmic audit standard aligned with the NIST AI Risk Management Framework, ensuring independent evaluation before deployment in high‑stakes settings. Governments could also implement a model procurement clause requiring vendors to disclose training data sources, document known limitations, and demonstrate mitigation of disparate impacts. Finally, Congress or state legislatures could establish statutory civil‑rights protections for automated decision systems, creating enforceable obligations regarding transparency, appeal rights, and safeguards against discrimination. These measures move beyond general guidance and create the structural conditions necessary for AI to advance opportunity rather than reinforce inequity.

 

References  

 

Egede, L. E. (2025, July 5). Exploring black communities’ perceptions and design approaches for building culturally tailored AI Systems | companion publication of the  2025 ACM Designing Interactive Systems Conference. ACM Digital Library. https:// dl.acm.org/doi/10.1145/3715668.3735629  

Elias, J. (2026, February 15). The Global South can shape AI in practical terms: Why the India AI Impact Summit Matters. OECD.AI. https://oecd.ai/en/wonk/the-global-south-can-shape-ai-in-practical-terms-why-the-india-ai-impact-summit-matters 

Finn, T., & Downie, A. (2025, November 17). Ai in recruiting. IBM.  https://www.ibm.com/think/topics/ai-in-recruitment  

Freedom House. (2024, February). Freedom in the world 2024. Freedom House. https://freedomhouse.org/sites/default/files/2024-02/FIW_2024_DigitalBooklet.pdf

Hargrove, S. K. (2024, December 5). A strategy for integrating artificial intelligence at historically Black Colleges & Universities. The Journal of Blacks in Higher Education. https://jbhe.com/2024/12/strategy-for-integrating-ai-at-hbcus/

Kalnina, V. (2025, November 6). AI governance at a crossroads: America’s AI Action Plan and its impact on businesses. Edmond & Lily Safra Center for Ethics. https:// www.ethics.harvard.edu/news/2025/11/ai-governance-crossroads-americas-ai action plan-and-its-impact-businesses  

NAACP. (2025, January 30). Ensuring representation and eliminating bias in artificial intelligence. NAACP.  

https://naacp.org/resources/ensuring-representation-and-eliminating-bias-artificial-int  elligence  

Radsch C., Lee N,(2024, December 23). Journalism needs better representation to counter  AI. Brookings.  

https://www.brookings.edu/articles/journalism-needs-better-representation-to-counter  ai/  

Swartsel, A. (2025, September 30). Unlocking the promise of AI for Black Learners and workers. Jobs for the Future (JFF).  

https://www.jff.org/idea/unlocking-the-promise-of-ai-for-black-learners-and-workers/