Economic Concentration, Political Influence, and Democratic Legitimacy in the AI Era

Full report

On this page

1. Longitudinal Trends in Public Trust, Perceptions of Corruption, and Institutional Responsiveness (1980–2026)

The fundamental stability of democratic governance rests upon political legitimacy: the widespread public conviction that political institutions possess the rightful authority to make and enforce binding collective decisions. Over the past four decades, public confidence in American political institutions has undergone a profound structural erosion. Long-term survey data compiled by the Pew Research Center, Gallup, and the National Election Studies demonstrate that while public trust in the federal government regularly exceeded 60% during the early 1960s, it collapsed during the Vietnam War and Watergate eras, achieved temporary partial recoveries during the mid-to-late 1980s and late 1990s, and subsequently entered a sustained downward trajectory1. Since 2004, the proportion of Americans who report trusting the government in Washington to do what is right "just about always" or "most of the time" has continuously remained below 40%5. By late 2025 and into October 2026, this metric dropped to near-historic lows ranging between 17% and 22%5. By October 2026, institutional confidence reached unprecedented depths across all three branches of the federal government8. Gallup tracking recorded public trust in the executive branch falling to 33%, the legislative branch to 28%, and the judicial branch to 46%—the lowest levels recorded for each respective branch8. Public confidence in the federal government’s capacity to handle domestic and international problems similarly stood at 33%8. Parallel to institutional decay, general social trust has fractured: only 51% of U.S. adults expressed trust in "the American people as a whole" to make political judgments, down from 83% in the early 1970s8.

A central driver of this institutional decay is the rising public perception of political corruption9. In September 2026, a record-high 89% of U.S. adults reported that government corruption was widespread throughout the federal government, representing a ten-percentage-point increase from 2025 and a substantial departure from the 72%–79% baseline observed between 2010 and 20259. Perceptions of corruption have largely transcended traditional partisan divisions: 91% of Democrats, 90% of Independents, and 83% of Republicans viewed government corruption as pervasive in late 20269. Perceptions of government corruption exceeded public perceptions of corporate corruption by 18 percentage points (89% versus 71%)9.

Comparative international metrics reflect a widening gap between the United States and peer OECD democracies9. In 2025, the gap in perceived government corruption between the U.S. (79%) and the OECD median (59%) expanded to 20 percentage points, with the U.S. tying Colombia for the highest perceived corruption rate among OECD member states9. Cross-national democratic health indicators demonstrate a corresponding trajectory10. The Varieties of Democracy (V-Dem) Institute’s Liberal Democracy Index (LDI)—which tracks electoral integrity, executive constraints, judicial independence, and civil liberties—recorded a drop in the U.S. score from 0.80 in 2015 down to 0.57 in 202510. In its 2026 report, V-Dem formally downgraded the U.S., placing its democracy score below all other Group of Seven (G7) nations and citing executive centralization and institutional friction10.

The systemic decline in public trust corresponds with a widespread perception of unresponsiveness6. Over 80% of U.S. adults consistently agree with the proposition that "elected officials don't care what people like me think"6. This disillusionment is tied to asymmetric trust patterns: citizens increasingly view federal authority as serving organized economic interests rather than the public interest5.

2. The Architecture of Corporate Political Influence in the Frontier Technology Sector

The accumulation of economic capital within the artificial intelligence sector has reshaped American campaign finance, lobbying operations, and administrative policy access. As capital expenditures for frontier AI development, advanced semiconductor manufacturing, and hyper-scale cloud infrastructure reached historic highs between 2023 and 2026, technology companies expanded their political operations in Washington D.C. and across state legislatures13. Federal Election Commission (FEC) records and disclosures filed under the Lobbying Disclosure Act (LDA) show that major technology companies spent over \$50 million on direct federal lobbying in the first three quarters of 2025 alone13. Between July and September 2025, combined lobbying outlays by dominant technology firms reached \$16 million—the highest single-quarter total recorded for the sector13. Alphabet led federal lobbying outlays with \$12.2 million over this period (an 11% year-over-year increase), while pure-play frontier AI developers registered the fastest rate of growth: OpenAI increased its quarterly lobbying expenditures by 68% year-over-year to \$2.1 million in Q3 2025, up from \$260,000 in full-year 202313. Simultaneously, specialized AI defense and intelligence contractors—including Palantir, Axon, and Parsons—tripled their federal lobbying expenditures between 2020 and 2025 to secure defense, border, and surveillance contracts14.

Beyond direct legislative lobbying, political influence in the technology sector relies on revolving-door recruitment—hiring former executive branch officials, congressional staffers, and regulatory agency attorneys15. Filings from the Senate Office of Public Records reveal that in 2023, 100% of OpenAI’s registered lobbyists (3 out of 3\) held prior government roles15. By 2024, as the firm expanded its government affairs team, 77.8% of its registered lobbyists (14 out of 18\) were former government personnel15. This recruitment strategy enables frontier firms to navigate complex legislative procedures, secure non-public access to regulatory decision-makers, and shape technical governance standards long before formal rulemaking notices are published13.

Complementing formal campaign contributions and lobbying disclosures, technology firms deploy multi-layered political influence strategies13. These include endowing academic chairs in computer science and technology policy, funding university research centers, sponsoring policy panels at major think tanks, and financing civil society coalitions focused on artificial intelligence ethics13. By underwriting academic research and policy scholarship, dominant technology companies shape the boundaries of regulatory debate13. Policy proposals that threaten corporate business models—such as public ownership of advanced compute clusters, structural unbundling of integrated cloud platforms, or mandatory open-sourcing of proprietary model weights—are frequently framed in policy discourse as risks to national competitiveness13.

3. Empirical Scholarship on Elite Influence, Interest Groups, and Policy Responsiveness

The central empirical question in political science regarding economic concentration is whether the policy preferences of wealthy citizens and organized corporate interests systematically outweigh those of average voters. Over the past two decades, political scientists have developed empirical frameworks to evaluate competing models of American democracy12.

Ideal-Type Theoretical Frameworks

Majoritarian Electoral Democracy posits that democratic elections force competing political parties to cater to the preferences of the median voter, ensuring that public policy aligns with majority public opinion12. In contrast, Economic-Elite Domination argues that policy outcomes are heavily skewed toward the preferences of high-income individuals and corporate elites, rendering mass public opinion largely non-influential when elite and public preferences diverge12. Biased Pluralism asserts that policy outcomes are dictated by organized interest groups, with business trade associations holding a structural advantage over labor unions and citizen coalitions12. Finally, Majoritarian Pluralism suggests that interest group competition broadly reflects mass public preferences, as membership organizations counteract corporate lobbies12.

The Empirical Dispute: Gilens, Page, and Methodological Re-Evaluations

In an influential empirical study analyzing 1,779 policy issues, Martin Gilens and Benjamin Page (2014) tested these four theories simultaneously using multivariate regression models12. Their central finding was that when the preferences of economic elites (measured at the 90th income percentile) and organized interest groups were held constant, the preferences of the average citizen (measured at the 50th income percentile) exerted a near-zero, statistically non-significant impact on public policy outcomes12. Gilens and Page concluded that the United States resembled a system of biased pluralism or oligarchy rather than a functional majoritarian democracy12. This conclusion generated significant debate within political science, prompting methodological re-examinations12. Subsequent scholarship challenged the oligarchy thesis on four main empirical grounds: First, Omar Bashir (2015) demonstrated that the policy preferences of the 50th and 90th income percentiles are highly correlated, frequently exceeding [Figure omitted from source export]12. Because median-income and high-income Americans agree on roughly 90% of policy proposals, multivariate regression models suffer from extreme multicollinearity12. Under conditions of high collinearity, standard linear regressions struggle to separate independent causal effects, making estimates sensitive to minor specification adjustments12. Second, J. Alexander Branham, Stuart N. Soroka, and Christopher Wlezien (2017) isolated the subset of policy issues (\~10%–15% of total cases) where high-income and median-income citizens explicitly disagreed (defined as a preference gap exceeding ten percentage points)21. In these contested policy disputes, the preferences of the affluent prevailed 53% of the time, while the preferences of the middle class prevailed 47% of the time22. Branham et al. argued that while the affluent hold a slight advantage, a 53-to-47 win rate contradicts claims of complete elite domination22. Third, research by Eric T. Brunner, Jeffrey R. Lax, and Justin H. Phillips evaluated Senate roll-call votes and found that political party affiliation remains a stronger predictor of legislative behavior than constituent wealth25. Their analysis showed that "party trumps the purse": legislators respond primarily to their primary voters and co-partisan constituents, meaning elite influence is largely mediated through partisan alignment25. Fourth, structural features of the U.S. political system—including the Senate filibuster, bicameralism, committee gatekeeping, and judicial review—create a systemic bias toward the political status quo26. While this status quo bias frequently prevents progressive redistributive policies favored by middle-and-low-income voters, it similarly blocks elite-backed legislative initiatives that lack broad political consensus26.

Synthesis of this empirical literature indicates that political power in the United States functions as a system of biased pluralism constrained by partisan polarization. Economic elites and well-capitalized technology corporations exert substantial influence over agenda-setting, regulatory drafting, and tax policy12. However, this influence is bounded by public opinion on highly visible issues, partisan competition, and institutional veto points25.

4. Stakeholder Political Advocacy across the AI Value Chain

The political economy of artificial intelligence spans a complex value chain: semiconductor designers, foundries, cloud providers, frontier model developers, application developers, labor unions, and civil society groups. Each tier of this value chain deploys distinct advocacy strategies to shape public policy13.

Compute, Hardware, and Semiconductor Manufacturers

At the physical foundation of the AI economy, hardware firms—led by graphics processing unit (GPU) designer Nvidia—focus political advocacy on export controls, supply chain subsidies, and antitrust scrutiny17. Facing Department of Justice antitrust investigations regarding its dominance in AI accelerators, semiconductor advocacy centers on framing hardware leadership as a national security asset17. These firms argue that aggressive antitrust interventions or strict export restrictions could erode American leadership in high-performance computing relative to international competitors18.

Cloud Providers and Hyperscalers

Integrated technology conglomerates operating cloud infrastructure—such as Microsoft (Azure), Amazon (AWS), and Alphabet (Google Cloud)—occupy a central position in the AI ecosystem13. Their political strategies concentrate on cloud licensing terms, federal procurement frameworks, energy access for data centers, and defending against tying and bundling allegations16. As major investors in frontier AI startups, cloud providers lobby to preserve strategic partnerships and cross-equity investments, resisting regulatory efforts to treat compute provisioning as a public utility16.

Frontier Developers and AI Startups

Firms developing foundation models—such as OpenAI, Anthropic, and Meta—exert legislative influence over AI safety rules, liability, and copyright regimes13. A significant policy fracture exists within this tier between proprietary developers and open-source advocates: Proprietary developers favor regulatory frameworks that mandate frontier safety audits, state or federal licensing regimes for models exceeding compute thresholds (e.g., [Figure omitted from source export] or [Figure omitted from source export]), and centralized oversight29. Critics contend these requirements create regulatory barriers that disadvantage smaller developers29. Conversely, open-source advocates and startups lobby against mandatory pre-deployment licensing, arguing that liability should attach to downstream misuse rather than upstream model development29. They advocate for open-weight models to democratize access and prevent market consolidation29.

Labor Organizations

Organized labor has mounted concerted counter-advocacy to mitigate technological displacement, protect worker autonomy, and secure compensation rights27. In October 2025, the AFL-CIO launched its "Workers First Initiative on AI," establishing policy principles that demand worker involvement in technology deployment, mandatory bargaining over workplace automation, and limits on algorithmic surveillance27. Creative and media unions have established contractual protections through collective bargaining31. The Writers Guild of America (WGA) negotiated terms establishing that generative AI cannot be classified as a "literary writer," ensuring that AI-generated drafts cannot reduce a human writer's compensation, credit, or residuals31. Similarly, SAG-AFTRA established mandatory bargaining protocols requiring explicit performer consent and compensation for the creation and deployment of "digital replicas" or synthetic performances31.

Civil Society and Public Research Advocates

Civil society coalitions, academic researchers, and public-interest groups focus on democratizing access to computing power and ensuring independent safety research29. A central objective has been the permanent legislative codification of the National Artificial Intelligence Research Resource (NAIRR)35. Through bipartisan legislation such as the CREATE AI Act (reintroduced in the 119th Congress as S. 4441 / H.R. 2385 by Senators Heinrich, Young, Rounds, and Booker), public-interest advocates seek to establish shared national compute infrastructure, cloud credits, and curated datasets managed by the National Science Foundation (NSF)35. This initiative aims to ensure that private technology companies do not hold a monopoly on advanced scientific research35.

5. Corporate Consolidation, Infrastructure Dependence, and Regulatory Policy

The capital requirements for training state-of-the-art foundation models have driven market consolidation across the AI value chain16. Developing frontier models requires multi-billion-dollar investments in specialized hardware, data center infrastructure, software optimization, and technical talent16. This capital intensity has fostered deep dependencies between emerging AI startups and dominant technology platforms16.

Federal Antitrust Interventions (2024–2026)

Recognizing these market concentration risks, federal antitrust enforcers—the Department of Justice (DOJ) Antitrust Division and the Federal Trade Commission (FTC)—divided oversight responsibilities to investigate potential anti-competitive conduct in the AI ecosystem41.

The Justice Department launched an investigation into Nvidia’s business practices, focusing on its 70% to 95% share of the global AI chip market17. Federal investigators issued subpoenas to examine whether Nvidia leveraged its market position by conditioning GPU allocations on whether customers purchased tied networking hardware (such as Quantum-2 InfiniBand switches) or software subscriptions17. Investigators are also evaluating whether Nvidia imposed price penalties on customers who attempted to deploy hardware from competing chipmakers, or utilized software acquisitions—such as RunAI—to make customer migration to non-Nvidia hardware more difficult17. Concurrently, the FTC expanded its investigation into Microsoft’s cloud computing and AI operations16. The probe focuses on whether Microsoft leveraged its dominance in enterprise software (Office 365, Windows, Entra ID) to favor its Azure cloud platform and Copilot tools while raising switching costs for rivals16. The FTC is also examining whether Microsoft utilized strategic equity investments and non-voting board positions in OpenAI to secure exclusive licensing and cloud hosting privileges while avoiding traditional Hart-Scott-Rodino merger review filings16.

Regulatory Capture, Preemption, and State-Level Battles

As federal antitrust agencies investigate structural concentration, state and federal legislative bodies have become sites of regulatory contestation29. A primary example was the debate over California’s Senate Bill 1047 (the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act) in 2024 and its subsequent replacement, Senate Bill 53, in September 202529. SB 1047 proposed safety mandates for developers training models above a [Figure omitted from source export] cost threshold, requiring pre-training safety protocols, mandatory third-party audits, and a "kill switch" capability to halt model operations in the event of catastrophic risk29. Following extensive lobbying by technology firms, venture capital firms, and industry alliances, Governor Gavin Newsom vetoed SB 1047 in late 202429. In its place, California enacted SB 53 (the Transparency in Frontier Artificial Intelligence Act) in September 202529. SB 53 shifted the regulatory framework from pre-deployment safety mandates and strict liability toward standardized public safety disclosures, internal risk-management frameworks, and transparency metrics29. At the federal level, technology industry advocacy groups have increasingly supported explicit federal preemption clauses44. Industry groups advocate for unified federal regulations that preempt state-level laws like California's SB 53 or New York's safety statutes44. Industry representatives argue that a patchwork of 50 distinct state regulatory standards creates compliance costs that harm innovation and national competitiveness44. Conversely, civil society organizations and state attorneys general contend that federal preemption, without robust federal standards, is designed to weaken state-level consumer protection and antitrust enforcement41.

6. Disentangling Institutional Influence, Systemic Distrust, Misconduct, and Allegations

To accurately evaluate democratic health in the AI era, analytical clarity requires separating four distinct political phenomena that are frequently conflated in public discourse. First, Verifiable Institutional Influence encompasses legally protected activities that tilt policy outcomes toward concentrated economic interests13. Examples backed by official documentation include over \$50 million in annual lobbying outlays by dominant technology companies, Super PAC spending utilizing non-disclosed "dark money" channels, revolving-door recruitment where over 77% of frontier AI firm lobbyists hold prior government roles, and legislative draft provisions submitted directly by corporate lobbyists to congressional committees13. Second, Public Perceptions of a "Rigged" System reflect broad public cynicism regarding institutional integrity5. Driven by economic inequality, persistent inflation, wage stagnation, and perceived government unresponsiveness, 89% of Americans view government corruption as widespread5. While these perceptions are anchored in structural inequalities, public frustration can obscure distinctions between legal political advocacy and illegal corruption5. Third, Documented Corporate Misconduct involves legal violations and regulatory infractions verified through administrative or judicial proceedings16. Examples include DOJ antitrust investigations establishing illegal tying practices or anti-competitive supply allocations in GPU markets, and enforcement actions penalizing firms for deceptive privacy practices or anti-competitive cloud licensing agreements16. Fourth, Unsupported Allegations of Election Manipulation comprise unsubstantiated claims alleging coordinated voter fraud, tampered voting machines, or illegitimate election outcomes10. Unlike documented campaign finance advantages or regulatory capture, these claims lack empirical verification11. However, when promoted by political figures, they amplify public cynicism, accelerate democratic backsliding, and erode trust in electoral processes10.

The interaction of economic exclusion, institutional distrust, and political unresponsiveness creates a reinforcing feedback loop. As technology-driven economic gains accumulate within a small tier of capital owners and specialized firms, citizens without capital ownership experience relative economic stagnation. Sensing that representative institutions are unresponsive to their economic needs, voters lose faith in democratic processes5. This institutional distrust leaves the public vulnerable to populist or autocratic narratives that promise to bypass traditional checks and balances6. When democratic norms weaken, institutional constraints on economic concentration erode further, reinforcing elite influence over the political economy11.

7. Comparative Models of Money in Politics and Technology Governance

The relationship between economic power and political responsiveness in the United States contrasts with regulatory frameworks in other advanced industrial democracies9. These comparative differences offer insight into how alternative political finance and governance models shape democratic legitimacy9.

Campaign Finance Frameworks

The American system treats political spending as protected speech under the First Amendment (rooted in Buckley v. Valeo and expanded in Citizens United v. FEC). This constitutional framework prevents the government from capping independent political expenditures by corporations, labor unions, or wealthy individuals. The resulting system relies on independent Super PACs, 527 organizations, and non-disclosed "dark money" channels, giving concentrated wealth significant influence over electoral campaigns. Conversely, European democracies regulate campaign finance using strict spending caps, restrictions or bans on paid political television advertising, shorter campaign windows, and robust public campaign funding. In France, corporate political donations are prohibited entirely, and individual campaign contributions are capped at low thresholds (€4,600 per election cycle). These mechanisms limit the direct influence of corporate wealth in election campaigns.

Regulatory Paradigms for Technology Governance

The U.S. approach to technology governance relies primarily on post-hoc, litigation-based market corrections and administrative enforcement through agencies like the FTC and DOJ16. While federal agencies can file antitrust suits against anti-competitive practices, litigation often takes years to resolve, during which market consolidation can become entrenched16. Legislative efforts often face parliamentary veto points and industry lobbying26. In contrast, the European Union relies on ex-ante, risk-based regulatory frameworks. The EU Artificial Intelligence Act categorizes AI applications into distinct risk tiers, imposing mandatory safety requirements, risk assessments, and technical auditing standards on high-risk and general-purpose AI models before market deployment. Similarly, the Digital Markets Act (DMA) designates large technology platforms as market "gatekeepers," imposing upfront obligations that prohibit self-preferencing, data bundling, and anti-competitive default settings16.

Comparative evidence indicates that nations with strict campaign spending limits, public party financing, and ex-ante regulatory frameworks maintain higher public trust in government institutions and lower rates of perceived corruption9.

8. Policy Reform Frameworks and Structural Tradeoffs

Addressing the challenges posed by economic concentration, political influence, and declining democratic legitimacy requires evaluating structural reform options. Each proposed policy intervention involves trade-offs among constitutional boundaries, administrative feasibility, economic innovation, and democratic representation.

Campaign Finance and Lobbying Transparency Reforms

Passing the DISCLOSE Act to eliminate non-disclosed dark money, establishing small-donor public matching funds for federal elections, and expanding mandatory lobbying disclosures to cover informal consulting can enhance political transparency13. Enacting a mandatory three-to-five-year cooling-off period for former executive branch officials and congressional staffers before they can engage in federal lobbying would mitigate revolving-door influence15. However, campaign spending restrictions face constitutional barriers under current Supreme Court First Amendment precedent. Furthermore, broader revolving-door restrictions could make public service less attractive to technical experts, potentially widening the expertise gap between private technology firms and public regulatory agencies15.

Antitrust Framework Evolution and Market Structure Interventions

Shifting from the narrow Consumer Welfare standard—which evaluates market power primarily through short-term consumer prices—toward Neo-Brandeisian frameworks considers market contestability, supply chain concentration, and structural political power16. Applying non-discrimination and anti-bundling rules to integrated cloud infrastructure providers and hardware manufacturers would prevent ecosystem leverage16. Nevertheless, aggressive antitrust enforcement and structural breakups could reduce operational efficiencies derived from vertically integrated cloud and compute systems16. Industry advocates warn that breaking up domestic technology leaders could weaken American technological leadership relative to state-subsidized foreign competitors18.

Public Options in Artificial Intelligence Compute (NAIRR)

Permanently codifying the National Artificial Intelligence Research Resource (NAIRR) through the passage of the CREATE AI Act (S. 4441 / H.R. 2385\) provides non-profit, academic, and civil society researchers with dedicated access to high-performance computing, open-source software environments, and curated datasets36. This democratizes compute access beyond dominant private firms35. However, building and maintaining competitive public compute infrastructure requires sustained federal funding38. If public compute resources lag behind private enterprise clusters, top-tier researchers may still migrate to private technology platforms35.

Ex-Ante Regulatory Frameworks and Federal Preemption Standards

Implementing federal safety disclosure and model transparency standards modeled after California’s SB 53 would require frontier developers to report risk assessment protocols, cybersecurity measures, and training compute footprints29. Yet broad federal preemption provisions risk invalidating state-level safety innovations and consumer protections44. Conversely, failing to preempt state laws subjects AI developers to conflicting regional compliance regimes, potentially raising entry barriers for smaller open-source developers29.

9. Conclusion: Empirical Certainties, Institutional Dilemmas, and Unresolved Questions

This investigation into economic concentration, political influence, and democratic legitimacy in the AI era reveals both established empirical trends and unresolved political-economic questions.

What Has Been Established Empirically

1. Historic Decline in Institutional Legitimacy: Public trust in U.S. government institutions has fallen to its lowest point in modern survey history (17%–20%), accompanied by a record 89% of citizens perceiving government corruption as widespread5. International comparative indices demonstrate that U.S. democratic standing has declined relative to OECD peer nations9.

2. Expansion of Political Outlays by the Technology Sector: As capital investment in frontier AI expanded, technology firms rapidly increased federal lobbying expenditures, exceeding \$50 million in the first three quarters of 2025 alone13.

3. High Revolving-Door Reliance: Leading AI firms rely heavily on former government officials for political representation, with frontier developers recording revolving-door lobbyist rates between 77% and 100%15.

4. Biased Pluralism in Legislative and Regulatory Outcomes: Empirical political science confirms that while public opinion can constrain politicians on high-visibility issues, economic elites and organized business groups exert outsized influence over technical rulemaking, tax policy, antitrust enforcement, and regulatory design12.

5. Structural Concentration Across the AI Value Chain: Market concentration in specialized hardware design (where Nvidia holds a 70%–95% share) and cloud infrastructure provision creates single points of leverage, prompting antitrust investigations by the DOJ and FTC16.

Unresolved Research Questions

1. Quantifying Negative Agenda Power: While political science can measure responsiveness to issues that reach a formal congressional vote, quantifying "negative agenda power"—the ability of concentrated capital to prevent policy options from entering public debate—remains an ongoing analytical challenge12.

2. Macroeconomic Dislocation and Democratic Stability: As generative AI and robotics advance, the extent to which technological labor displacement will accelerate democratic backsliding versus triggering social safety responses (such as universal basic income or public compute dividends) remains untested empirically10.

3. Efficacy of Post-Hoc Antitrust Enforcement: It remains uncertain whether traditional antitrust remedies (such as civil investigative demands, consent decrees, or structural breakups) can operate quickly enough to prevent market consolidation in rapidly evolving technology markets16.

4. Counterbalancing Private Compute Dominance: Whether public infrastructure initiatives like the National Artificial Intelligence Research Resource (NAIRR) can provide sufficient compute scale to preserve independent academic research alongside private technology platforms remains an open question35.

In summary, as artificial intelligence increases the economic value of technology ownership, the central challenge for democratic governance is preventing concentrated private economic power from converting into unmatchable political influence12. Restoring public confidence in representative institutions will depend on whether democratic governments can maintain regulatory independence, preserve market contestability, and ensure that technological progress reinforces broad democratic legitimacy6.

Works cited

1. Public trust in the United States government, 1958 to 2025, https\://ourworldindata.org/grapher/public-trust-in-government

2. Public Trust in Government: 1958-2025 \- Pew Research Center, https\://www\.pewresearch.org/politics/2025/12/04/public-trust-in-government-1958-2025/

3. Public trust in government near historic lows \- Pew Research Center, https\://www\.pewresearch.org/chart/public-trust-in-government-near-historic-lows-5/

4. Americans' Deepening Mistrust of Institutions, https\://www\.pew.org/en/trend/archive/fall-2024/americans-deepening-mistrust-of-institutions

5. The State of Public Trust in Government 2026, https\://ourpublicservice.org/know-the-facts/resource-library/reports/state-of-trust-2026

6. Trust in Government \- Research and data from Pew Research Center, https\://www\.pewresearch.org/topic/politics-policy/trust-in-government/

7. Pew report shows Americans' trust in government at historic lows, https\://www\.youtube.com/watch?v=H8cTy19OO6g

8. Trust in Federal Government Sinks to New Depths \- Gallup News, https\://news.gallup.com/poll/715361/trust-federal-government-sinks-new-depths.aspx

9. Record-High 89% in U.S. Say Government Corruption Widespread, https\://news.gallup.com/poll/713933/record-high-say-government-corruption-widespread.aspx

10. Health of US democracy in decline, based on multiple evaluations, https\://www\.pewresearch.org/short-reads/2026/04/15/multiple-indicators-show-a-decline-in-the-health-of-americas-democracy-in-2025/

11. V-DEM Democracy Report 2026, https\://www\.v-dem.net/documents/75/V-Dem\_Institute\_Democracy\_Report\_2026\_lowres.pdf

12. (PDF) Testing Inferences about American Politics: A Review of the, https\://www\.researchgate.net/publication/283035709\_Testing\_Inferences\_about\_American\_Politics\_A\_Review\_of\_the\_Oligarchy\_Result

13. As Big Tech Gears Up for the 2026 Midterms, Its Lobbying, https\://issueone.org/articles/big-tech-lobbying-2025-q3/

14. Tech firms triple lobbying expenditures while raking in millions from, https\://www\.opensecrets.org/news/2026/04/palantir-axon-parsons-triple-lobbying-expenditures-while-raking-in-millions-from-ice-contracts/

15. OpenAI Profile: Summary \- OpenSecrets, https\://www\.opensecrets.org/orgs/openai/summary?id=D000084252

16. Microsoft Antitrust Probe Widens: 6 Rivals Queried \[2026\], https\://tech-insider.org/microsoft-antitrust-probe-cloud-ai-2026/

17. NVIDIA Hit with DOJ Antitrust Probe over AI GPUs, Unfair Sales, https\://www\.techpowerup.com/forums/threads/nvidia-hit-with-doj-antitrust-probe-over-ai-gpus-unfair-sales-tactics-and-pricing-alleged.325160/

18. Nvidia Faces DOJ Antitrust Investigation: A Potential Turning Point In, https\://www\.legal500.com/intelligence/united-states/antitrust-competition-law/nvidia-faces-doj-antitrust-investigation-a-potential-turning-point-in-the-ai-industry

19. The DOJ and Nvidia: AI Market Dominance and Antitrust Concerns, https\://www\.americanactionforum.org/insight/the-doj-and-nvidia-ai-market-dominance-and-antitrust-concerns/

20. Testing Theories of American Politics: Elites, Interest Groups, and, https\://www\.researchgate.net/publication/278151684\_Testing\_Theories\_of\_American\_Politics\_Elites\_Interest\_Groups\_and\_Average\_Citizens

21. Analyze inequality and representation links \- Varsity Tutors, https\://www\.varsitytutors.com/practice/subjects/college-political-science/lessons/inequality-and-representation

22. Economic Inequality and Political Responsiveness: A Systematic, https\://www\.cambridge.org/core/journals/perspectives-on-politics/article/economic-inequality-and-political-responsiveness-a-systematic-review/3364318C95A3D608048BA1800013C7E1

23. What to Know Before Citing Gilens and Page (2014), “Testing, https\://medium.com/@omar.bashir/what-to-know-before-citing-gilens-and-page-2014-testing-theories-of-american-politics-8918e69ba0d7

24. BACKGROUNDER \- The Heritage Foundation, https\://www\.heritage.org/sites/default/files/2017-06/BG3227.pdf

25. The Party or the Purse? Unequal Representation in the US Senate, https\://www\.columbia.edu/\~jrl2124/partypurse.pdf

26. Who got what they wanted? Investigating the role of institutional, https\://www\.tandfonline.com/doi/full/10.1080/13501763.2023.2177323

27. AI and Labor \- AFL-CIO, https\://aflcio.org/issues/future-work/ai

28. US antitrust enforcers to investigate AI companies Microsoft, Nvidia, https\://apnews.com/article/nvidia-openai-microsoft-ai-antitrust-investigation-ftc-doj-0adc9a4a30d4b581a4f07894473ba548

29. SB 53: What California's New AI Safety Law Means for Developers, https\://ai-analytics.wharton.upenn.edu/wharton-accountable-ai-lab/sb-53-what-californias-new-ai-safety-law-means-for-developers/

30. California adopts landmark AI law: new transparency, safety, and, https\://www\.aoshearman.com/en/insights/ao-shearman-on-tech/california-adopts-landmark-ai-law

31. Hollywood Unions Take Different Approaches to AI Protections, https\://journals.law.harvard.edu/jsel/2023/10/hollywood-unions-take-different-approaches-to-ai-protections/

32. SAG-AFTRA Gives Presentation on AI in the Workplace at the AFL, https\://www\.sagaftra.org/sag-aftra-gives-presentation-ai-workplace-afl-cio-convention

33. SAG-AFTRA A.I. Bargaining And Policy Work Timeline, https\://www\.sagaftra.org/sag-aftra-ai-bargaining-and-policy-work-timeline

34. How Unions Can Advance AI Protections Across Sectors | SAG-AFTRA, https\://www\.sagaftra.org/negotiating-ai-contracts-how-unions-can-advance-ai-protections-across-sectors

35. Senators take another swing at bill to codify federal AI resource, https\://fedscoop.com/create-ai-act-senate-bill-nairr/

36. Heinrich, Young, Rounds, Booker Reintroduce Bipartisan, https\://www\.heinrich.senate.gov/newsroom/press-releases/heinrich-young-rounds-booker-reintroduce-bipartisan-legislation-to-expand-access-to-artificial-intelligence-research

37. Bipartisan Senators Renew Their Push to Codify NAIRR \- MeriTalk, https\://www\.meritalk.com/articles/bipartisan-senators-renew-their-push-to-codify-nairr/

38. the NAIRR Operations Center (NAIRR-OC) | NSF \- U.S. National, https\://www\.nsf.gov/funding/opportunities/nairr-oc-foundations-operating-national-artificial-intelligence/506437/nsf25-546

39. NSF 25-546: Foundations for Operating the National Artificial, https\://www\.nsf.gov/funding/opportunities/nairr-oc-foundations-operating-national-artificial-intelligence/506437/nsf25-546/solicitation

40. National AI Research Resource (NAIRR) (2026) \- Granted AI, https\://grantedai.com/grants/national-ai-research-resource-nairr-national-science-foundation-nsf-in-partn-e5cf18d6

41. DOJ, FTC Target AI Industry In Antitrust Investigation \- Legal 500, https\://www\.legal500.com/en/intelligence/united-states/antitrust-competition-law/doj-ftc-target-ai-industry-in-antitrust-investigation-three-takeaways

42. Antitrust Investigations into Microsoft, OpenAI, and NVIDIA, https\://cyberjustice.blog/2024/09/02/antitrust-investigations-into-microsoft-openai-and-nvidia/

43. Microsoft, OpenAI and Nvidia investigated over monopoly laws, https\://www\.theguardian.com/business/article/2024/jun/06/microsoft-openai-and-nvidia-investigated-over-possible-breach-of-antitrust-laws

44. When Should Congress Preempt State AI Law? The Lessons of Past, https\://carnegieendowment.org/research/2025/09/congress-preempt-state-ai-law-the-lessons-of-past-technologies

45. Governor Newsom announces results from California's first-in-the, https\://www\.gov.ca.gov/2026/10/13/governor-newsom-announces-results-from-californias-first-in-the-nation-project-to-involve-californians-in-decisions-about-ai/

46. Federal Preemption of AI Regulation: What State Legislation Is at, https\://ari.us/wp-content/uploads/2025/06/ARI-Report-Fed-Preemption-6-6-2025.pdf

47. Summary of Artificial Intelligence 2025 Legislation, https\://www\.ncsl.org/technology-and-communication/artificial-intelligence-2025-legislation

48. AI Companies Targeted in New FTC & DOJ Antitrust Investigation, https\://www\.openmarketsinstitute.org/publications/ai-companies-targeted-ftc-doj

49. CREATE AI Act of 2025 creates a National AI Research Resource, https\://codifylegalpublishing.com/blog-article?id=bill-us-united-states-119th-hb2385-create-ai-2025-resources-american-experiment