The Information Machine

US AI Regulation: Federal Retreat vs. State Intervention · history

Version 22

2026-06-21 02:16 UTC · 324 items

What

The Trump administration maintains voluntary federal AI oversight while exercising direct national security enforcement — forcing Anthropic to disable its most advanced models for foreign nationals [6] and protecting xAI infrastructure [7] — while opposing binding domestic safety requirements. State governments are advancing independently: Illinois's enacted AI safety law [9], California's pending employment bills [10][11], and a 42-state coalition subpoena targeting OpenAI [12] are all progressing without federal coordination. VP Vance joined David Sacks in publicly framing AI safety advocacy as fear-based marketing [20][19], while Anthropic CEO Amodei defends mandatory oversight proposals — tensions that sit awkwardly against the confirmed government enforcement action against Anthropic itself. Open-source AI advocates warn this regulatory landscape risks consolidating market power between Anthropic and OpenAI at the expense of open-weight model alternatives [27].

Why it matters

The administration simultaneously restricts Anthropic's products on national security grounds and dismisses Anthropic's safety arguments as self-interested fear-mongering — an internally inconsistent posture that neither the tech accelerationist camp nor Anthropic can fully account for. State enforcement continues to fill the regulatory vacuum the federal government is deliberately leaving, producing a fragmented but binding de facto framework regardless of federal preferences.

Open questions

  • Will the Anthropic export control action expand to other frontier labs, or is it specific to Anthropic's technology or customer base? [6][28]

  • Does VP Vance's alignment with Sacks on fear-mongering [20] represent an administration-wide policy stance, or is it personal commentary without formal policy consequence?

  • Will Lambert's open-source AI arguments [27] gain policy traction, given the administration's national security framing that could justify restricting open-weight models as readily as proprietary ones?

  • Will the 42-state coalition subpoena against OpenAI [12] and Florida's expanded lawsuit [14] produce coordinated settlement terms, or will they remain parallel pressure tracks?

Narrative

Trump signed an AI executive order on June 3, 2026, establishing a voluntary 30-day pre-release review for frontier models with NSA oversight and classified compliance thresholds, explicitly disclaiming binding requirements [1][2][3]. The order's weakened form reflects lobbying by Musk, Zuckerberg, and David Sacks, who stripped binding requirements and cut the review window from 90 to 30 days [4][5]. Alongside the voluntary EO, the administration exercises direct enforcement on national security grounds: the US government forced Anthropic to disable its most advanced models for all foreign nationals [6], and the DOJ protected xAI's Mississippi data center as national security infrastructure [7]. The Pentagon has deployed Google Cloud's Gemini to all six military branches and its CTO publicly cited AI-drafted congressional reports — produced in five hours rather than 200 staff-hours — as a flagship adoption example [8].

State governments are advancing binding AI law and enforcement independent of federal coordination. Illinois enacted SB 315 requiring annual third-party audits, 72-hour incident reporting, and whistleblower protections [9]. California is advancing SB 947 requiring human oversight of automated employer discipline and SB 951 requiring 90-day notice of AI-driven layoffs [10][11]. On June 12, 2026, New York AG Letitia James served OpenAI with a subpoena on behalf of a 42-state coalition targeting advertising practices, data handling, treatment of minors, and chatbot sycophancy [12]. Florida filed an 83-page lawsuit naming CEO Sam Altman personally, growing from a criminal investigation tied to a Florida State University shooting [13][14]. Trump's preemption executive order challenged state AI laws on constitutional grounds [15], but multiple law firms found its authority limited absent congressional action [16].

The most direct public argument over federal AI regulation runs between Anthropic CEO Dario Amodei and the tech accelerationist wing of the administration. Amodei published a blog calling for mandatory pre-release testing and government-backed independent auditing [17][18]. Sacks responded by accusing Anthropic of 'a sophisticated regulatory capture strategy based on fear-mongering' [19], and VP Vance publicly echoed this position, arguing AI CEOs promote dystopian narratives because fear makes their products appear more capable and valuable [20]. Amodei disputed the characterization, claiming alignment with Trump administration priorities [21]. The administration's confirmed export control action against Anthropic [6] sits uneasily against both arguments: Sacks serves an administration that has restricted Anthropic's products, while Amodei claims alignment with an administration that has treated his company as a national security concern. OpenAI publicly advocates for federal governance [22][23] while its affiliated PAC Build American AI — backed by OpenAI, Andreessen/a16z, and Palantir — admitted running sockpuppet accounts and paid TikTok influencers to fear-monger about Chinese AI [24][25][26]; the 42-state subpoena directly targets the consumer practices OpenAI frames as responsible [12]. From outside the major labs, Nathan Lambert argues that Anthropic and OpenAI are forming a closed-model duopoly and that regulatory focus on frontier models risks inadvertently restricting open-source AI, further concentrating market power between these two companies [27].

Timeline

  • 2025-12: Trump signs executive order aimed at eliminating state law obstruction of national AI policy [29][55][56]
  • 2026-01-01: New state AI laws take effect nationally, triggering the federal preemption debate [57]
  • 2026-05: Trump signs preemption EO challenging state AI laws; multiple law firms find constitutional authority limited absent congressional authorization [15][55][16][58][59]
  • 2026-05: California Senate advances SB 947 requiring human oversight of automated employer discipline and SB 951 requiring 90-day advance notice of AI-driven layoffs [11][42][10][43]
  • 2026-05-21: Trump cancels AI security EO after Sacks calls Trump directly to oppose pre-deployment review; Musk and Zuckerberg also lobby against binding requirements [4][5][60]
  • 2026-05-28: Illinois Governor Pritzker signs SB 315, the nation's strictest AI safety statute, requiring mandatory third-party auditing, 72-hour critical incident reporting, and whistleblower protections [44][45][9]
  • 2026-06-01: Florida becomes first state to sue OpenAI and Sam Altman over ChatGPT's alleged role in violent crimes; AG Uthmeier also opens a separate criminal investigation [46][47][13]
  • 2026-06-03: Trump signs revised AI executive order establishing a voluntary 30-day pre-release review for frontier models with NSA oversight and classified compliance thresholds [1][2][30][3]
  • 2026-06-03: OpenAI publishes Frontier Governance Framework calling for coherent federal architecture over fragmented state regulation [37][22][38][39]
  • 2026-06-04: Build American AI — backed by OpenAI, Andreessen/a16z, and Palantir — acknowledged running false flag social media accounts; WIRED reports the PAC paid TikTok influencers to fear-monger about Chinese AI [24][25][40][26]
  • 2026-06-10: Anthropic CEO Dario Amodei publishes blog calling for mandatory pre-release testing and government-backed independent auditing [17][18]
  • 2026-06-12: New York AG Letitia James serves OpenAI with a subpoena on behalf of a 42-state coalition, targeting advertising practices, data handling, treatment of minors, and chatbot sycophancy [12]
  • 2026-06-14: David Sacks publicly accuses Anthropic of 'regulatory capture strategy based on fear-mongering'; Amodei responds, defending Anthropic's regulatory stance as aligned with Trump priorities [31][19][33][21][34]
  • 2026-06-14: US government forces Anthropic to disable its most advanced AI models for all foreign nationals; Just Security publishes legal analysis of the action [6][36]
  • 2026-06-15: Florida expands its lawsuit against OpenAI to 83 pages naming CEO Sam Altman personally, separate from the ongoing criminal probe [12][14][48]
  • 2026-06-16: Pentagon CTO Emil Michael cites AI-drafted congressional reports — produced in five hours instead of 200 staff-hours — as a flagship DoD AI adoption example at a Hudson Institute event [8]
  • 2026-06-18: DOJ moves to protect xAI's Mississippi AI data center, treating high-speed AI compute as national security infrastructure [7]
  • 2026-06-18: Reporting indicates Sacks distinguishes a formal export-control action against Anthropic from the public regulatory dispute and expects resolution [35]
  • 2026-06-19: VP Vance publicly accuses AI CEOs of using fear as a marketing strategy, echoing Sacks' critique of AI safety advocacy [20]
  • 2026-06-19: Nathan Lambert publishes argument against open-source AI regulation, warning executive actions targeting frontier models could concentrate market power at Anthropic and OpenAI [27]

Perspectives

Trump administration (federal)

Signed a voluntary AI EO with NSA oversight and 30-day review window explicitly disclaiming binding requirements; simultaneously forced Anthropic to disable models for foreign nationals, directed DOJ to protect xAI's data center as national security infrastructure, and deployed AI across the Pentagon to draft congressional oversight documents.

Evolution: Administration posture has expanded from advisory review (EO) to direct enforcement (Anthropic export controls), asset protection (xAI/DOJ), and active government AI adoption (Pentagon), while VP Vance has aligned publicly with Sacks' anti-regulation rhetoric.

[29][15][4][1][2][30][3][7][6][8][20]

Tech accelerationist faction (Sacks, Musk, Zuckerberg, Vance)

Opposes AI oversight mechanisms; Sacks accuses Anthropic of 'a sophisticated regulatory capture strategy based on fear-mongering'; VP Vance echoes the same framing, arguing AI CEOs promote dystopian narratives because fear makes their products appear more capable.

Evolution: VP Vance's public alignment with Sacks' critique expands the faction from tech industry figures to sitting administration leadership, while the administration they are part of has confirmed the export control action against Anthropic.

[4][5][31][19][20]

Anthropic / Dario Amodei

Calls for mandatory pre-release testing and government-backed independent auditing; disputes Sacks' and Vance's 'fear-mongering' characterization while claiming alignment with Trump administration priorities.

Evolution: Now confirmed subject of a government export control action requiring it to disable most advanced models for foreign nationals, while simultaneously defending its regulatory advocacy as pro-administration — a difficult dual position given Vance's and Sacks' public attacks.

[32][18][17][33][21][34][35][6][36]

OpenAI

Publicly advocates for federal governance and equitable industrial policy while its affiliated PAC admitted running sockpuppet accounts and paying influencers; facing a 42-state coalition subpoena targeting the advertising and consumer practices the company frames as responsible.

Evolution: The 42-state subpoena directly contradicts OpenAI's governance advocacy by targeting consumer-facing practices the company frames as responsible.

[37][22][38][39][24][25][40][26][23][41][12]

State governors advancing AI legislation (Newsom + Pritzker)

State government must proactively address AI harms; Illinois SB 315 enacted as the nation's strictest AI safety statute; California advancing SB 947 and SB 951 on employment-related AI decisions.

Evolution: Consistent; the two-state legislative coalition operates independently of federal coordination.

[11][42][10][43][44][45][9]

State AGs (42-state coalition + Florida litigation)

Pursues AI accountability through civil and criminal enforcement; the 42-state coalition subpoenaed OpenAI over advertising, data handling, and chatbot behavior; Florida filed an 83-page suit naming Altman personally.

Evolution: Expanded from Florida's single-state lawsuit to a 42-state coordinated investigation — the first major probe targeting how a chatbot behaves, not just what it does with data.

[46][47][13][12][14][48]

Congress — federal legislative responses

Activity remains fragmented: Sanders proposed a 50% stock transfer to a public sovereign fund; Senator Warren pressed Nvidia on chip exports; no binding legislation is advancing.

Evolution: Consistent; congressional activity remains fragmented relative to the pace of state enforcement.

[49][50][51][52][53][54]

Nathan Lambert / open-source AI advocates

Argues that restricting open-source AI would harm startups, universities, and enterprises that depend on open-weight models as affordable alternatives; characterizes Anthropic and OpenAI as a 'duopoly' concentrating market power; warns that regulatory focus on frontier models could push global users toward Chinese open-source alternatives.

Evolution: New voice in this synthesis; enters as the regulatory debate expands beyond the Amodei-Sacks axis to include questions of market concentration and open-source access.

[27]

Tensions

  • Mandatory oversight (Amodei) vs. accelerationist opposition (Sacks, Vance): Amodei calls for mandatory pre-release testing and government-backed auditing [17]; Sacks and VP Vance both argue AI safety advocates manufacture fear for commercial reasons [19][20] — yet the administration they serve has forced Anthropic to disable its most advanced models for foreign nationals [6]. [4][5][17][31][19][6][20]
  • Federal preemption preference vs. state enforcement reality: the Trump preemption EO lacks constitutional authority absent congressional action [15][16]; binding Illinois law [9] and a 42-state coalition subpoena [12] advance without federal coordination. [15][16][9][38][39][12]
  • Federal enforcement of national security AI restrictions vs. resistance to domestic consumer safety requirements: the government forced Anthropic to disable models for foreign nationals [6] and protected xAI's data center as national security infrastructure [7], while maintaining that binding domestic safety requirements are unnecessary. [7][6][1][3]
  • OpenAI's governance advocacy vs. its PAC's political operations and legal exposure: OpenAI calls for democratic governance [22][23] while Build American AI admitted running sockpuppet accounts [41] and the 42-state subpoena targets the consumer practices OpenAI frames as responsible [12]. [22][23][24][41][12]
  • Anthropic's dual positioning: Amodei claims mandatory oversight aligns with Trump administration priorities [21], while Sacks argues Anthropic is principally responsible for the 'state regulatory frenzy' [19] — both claims cannot hold simultaneously, and the confirmed export control action against Anthropic [6] complicates both. [21][19][6]
  • Open-source AI advocates vs. closed-model incumbents: Lambert argues regulatory focus on frontier models could inadvertently restrict open-source AI and concentrate market power at Anthropic and OpenAI [27]; Anthropic argues mandatory oversight is a safety necessity, not a competitive strategy [17][21]. [27][17][21]

Sources

  1. [1] Trump Signs Executive Order For AI Testing Prior To Frontier Model Releases — Zvi's AI Roundups (2026-06-03)
  2. [2] President Trump signs executive order to review advanced AI models. — Rohan Paul Twitter (2026-06-03)
  3. [3] Executive Order Creates Voluntary Regulatory Regime of Frontier AI ... — reactive:us-ai-policy-regulation
  4. [4] Trump abruptly cancels EO signing event after top AI firm CEOs declined to go — Ars Technica AI (2026-05-22)
  5. [5] AI #170: Lack of Executive Order — Zvi's AI Roundups (2026-05-28)
  6. [6] The US government just forced Anthropic to disable its most advanced AI models for all foreign nationals — and the compa... — reactive:us-ai-policy-regulation (2026-06-14)
  7. [7] Large AI campuses are indeed strategic infrastructure of a nation. — Rohan Paul Twitter (2026-06-18)
  8. [8] Pentagon boasts of using AI to write reports mandated by Congress — Ars Technica AI (2026-06-16)
  9. [9] Illinois Signs Frontier AI Safety Bill into Law | Chris Lehane posted ... — reactive:us-ai-policy-regulation
  10. [10] CA Senate Approves No Robo Bosses Act of 2026 to Ensure Human Oversight of AI in the Workplace | Senator Jerry McNerney — reactive:us-ai-policy-regulation
  11. [11] Bill Text: CA SB951 | 2025-2026 | Regular Session | Amended — reactive:us-ai-policy-regulation
  12. [12] 😺 42 states just subpoenaed OpenAI — The Neuron (2026-06-15)
  13. [13] Florida AG brings 'first-in-the-nation state-led lawsuit' against OpenAI and its CEO — reactive:us-ai-policy-regulation
  14. [14] Florida AG Sues OpenAI, Alleges ChatGPT Aided Mass Shooters — reactive:us-ai-policy-regulation
  15. [15] President Trump Signs Executive Order Preempting State AI Laws ... — reactive:us-ai-policy-regulation
  16. [16] President Trump Signs Executive Order Challenging State AI Laws | Paul Hastings LLP — reactive:us-ai-policy-regulation
  17. [17] Dario Amodei just published a super long blog, calling for an urgent policy overhaul because he thinks frontier AI is mo… — Rohan Paul Twitter (2026-06-10)
  18. [18] Dario Amodei argues in his new blog today, that powerful AI will soon become the dominant source of military and economi… — Rohan Paul Twitter (2026-06-11)
  19. [19] Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering. It is principally responsible for the state regulatory frenzy that is damaging the startup ecosystem. — reactive:us-ai-policy-regulation
  20. [20] US VP JD Vance on AI Company's marketing strategy, some AI CEOs sell fear because fear makes their product look powerful… — Rohan Paul Twitter (2026-06-19)
  21. [21] Anthropic CEO defends support for AI regulations, alignment with ... — reactive:us-ai-policy-regulation
  22. [22] OpenAI public policy agenda — OpenAI Blog (2026-06-03)
  23. [23] Industrial policy for the Intelligence Age — OpenAI Blog (2026-06-09)
  24. [24] AI #171: False Flag — Zvi's AI Roundups (2026-06-04)
  25. [25] The OpenAI-Andreessen-Palantir SuperPAC admits that it was “part ... — reactive:us-ai-policy-regulation
  26. [26] A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat | WIRED — reactive:us-ai-policy-regulation
  27. [27] Banning Open Source AI Would Be A Mistake — Interconnects (2026-06-19)
  28. [28] The Day Silicon Valley's AI Crown Nearly Shattered: Inside Anthropic's Brush With National Security Prosecution — reactive:us-ai-policy-regulation (2026-06-20)
  29. [29] Ensuring a National Policy Framework for Artificial Intelligence — reactive:us-ai-policy-regulation
  30. [30] Trump AI Executive Order Gives NSA Classified Role Over Frontier Models — reactive:us-ai-policy-regulation
  31. [31] David Sacks is done being polite about Anthropic (Save this). — Milk Road AI Twitter (2026-06-14)
  32. [32] Results from the first Anthropic Public Record — Anthropic News (2026-06-12)
  33. [33] Anthropic CEO disputes David Sacks' claims that company is 'woke' — reactive:us-ai-policy-regulation
  34. [34] Anthropic CEO claps back after Trump officials accuse firm of AI fear-mongering   | TechCrunch — reactive:us-ai-policy-regulation
  35. [35] Sacks separates Anthropic export-control action from ongoing legal dispute, expects resolution | InsideAIPolicy.com — reactive:us-ai-policy-regulation
  36. [36] Legal Considerations Related to the Anthropic “Export Controls ... — reactive:us-ai-policy-regulation
  37. [37] OpenAI’s Frontier Governance Framework — OpenAI Blog (2026-05-28)
  38. [38] A blueprint for democratic governance of frontier AI — OpenAI Blog (2026-06-03)
  39. [39] 🟡 Let’s get physical — Semafor Technology (2026-06-03)
  40. [40] A Pro-AI Super PAC’s Secret Meme Sockpuppets — reactive:us-ai-policy-regulation
  41. [41] Build American AI admits hiring an outside vendor to run anonymous sockpuppet accounts targeting its opponents · Digg — reactive:us-ai-policy-regulation
  42. [42] SB 951: Employment: technological displacement: notice. — reactive:us-ai-policy-regulation
  43. [43] California Set to Restrict AI Use in the Workplace With “No Robo ... — reactive:us-ai-policy-regulation
  44. [44] Trump loses more control over AI regulation as Illinois passes landmark law — Ars Technica AI (2026-05-28)
  45. [45] Illinois Legislature passes historic AI bill that would require third-party safety audits — reactive:us-ai-policy-regulation
  46. [46] ChatGPT creators knew product would cause harm, Florida argues in lawsuit • Florida Phoenix — reactive:us-ai-policy-regulation
  47. [47] Attorney General James Uthmeier Launches Criminal Investigation into OpenAI, ChatGPT | My Florida Legal — reactive:us-ai-policy-regulation
  48. [48] Florida Attorney General Sues OpenAI Over ChatGPT's Alleged ... — reactive:us-ai-policy-regulation
  49. [49] Text - S.3339 - 119th Congress (2025-2026): AI Workforce PREPARE Act — reactive:us-ai-policy-regulation
  50. [50] Rep. Obernolte, Rep. Jacobs Introduce Bipartisan Bill to Prepare American Workers for AI-Driven Economic Change | Representative Jay Obernolte — reactive:us-ai-policy-regulation
  51. [51] Sanders Sovereign Wealth Fund Plan Would Give US Public 'Direct Ownership Stake' in AI Giants | Common Dreams — reactive:us-ai-policy-regulation
  52. [52] Sen. Bernie Sanders on X: "AI is built on humanity’s collective knowledge. The wealth it generates must benefit humanity — not just Elon Musk, Sam Altman and other AI oligarchs. That’s why I’ll be introducing the American AI Sovereign Wealth Fund Act — to give the public a direct ownership stake. https://t.co/UqW71FBv2Z" / X — reactive:us-ai-policy-regulation
  53. [53] 💯💯 David Sacks rips Bernie Sanders' 50% AI ownership plan https://t.co/hPsHZ6eOyJ https://t.co/yD9Ko0V6oP — Rohan Paul Twitter (2026-06-07)
  54. [54] Jensen Huang just refused Sen. Elizabeth Warren’s request to testify before Congress on AI chips, China sales, and expor… — Rohan Paul Twitter (2026-06-08)
  55. [55] President Trump Signs Executive Order to Block State AI Laws — reactive:us-ai-policy-regulation
  56. [56] Executive Order: Ensuring a National Policy Framework for Artificial Intelligence (Donald Trump, 2025) - Ballotpedia — reactive:us-ai-policy-regulation
  57. [57] New State AI Laws are Effective on January 1, 2026, But a New ... — reactive:us-ai-policy-regulation
  58. [58] When Federal Preemption Meets AI Regulation: What Trump's Draft Executive Order Means for Your Compliance Strategy | Jones Walker LLP — reactive:us-ai-policy-regulation
  59. [59] Executive Preemption and the Dormant Commerce Clause After ... — reactive:us-ai-policy-regulation
  60. [60] WATCH: Trump explains why he postponed signing AI executive order — reactive:us-ai-policy-regulation