Event Brief
Sen. Bernie Sanders and Rep. Greg Casar unveiled the Ban Artificial Superintelligence Act on September 23, 2026, with the bill text provided first to the Associated Press. The legislation would permanently prohibit the development and deployment of artificial superintelligence while temporarily pausing advanced AI development generally until a new federal regulator establishes safety rules and a model-review process. The bill would create a new Cabinet-level Department of Artificial Intelligence, advised by an Artificial Intelligence Advisory Board of technical experts, tasked with monitoring frontier AI systems across their lifecycle, supervising removal of dangerous capabilities, and supervising outright destruction of any superintelligent system found to exist.
The enforcement mechanism is unusually severe for tech policy. Per the sponsors' own release summary, entities that violate the pauses and prohibitions face a "corporate death penalty," while individuals face up to 20 years in prison — a penalty the bill's backers explicitly liken to existing statutory penalties for unlawfully developing nuclear weapons. The bill also directs U.S. foreign policy toward negotiating international agreements, allied coordination, and export controls to prevent superintelligence development anywhere in the world, extending the ban's ambition beyond U.S. jurisdiction.
The bill's release lands in direct tension with the executive branch's posture. On September 19-20, 2026, the President announced via social media that he would form an "AI Force" modeled on the Space Force and name a new AI "czar," while dismissing fears of runaway superintelligence as a "hoax" and a "SICK conspiracy," and vowing the administration would "not in any way hinder or stifle" AI industry growth. House Speaker Mike Johnson has said the president is not interested in new federal AI regulation, though he has reportedly agreed to meet with AI industry leaders. This sets up a direct substantive collision between the Sanders-Casar approach (prohibition, criminal liability, forced dissolution) and the administration's approach (self-policing via existing criminal/civil law, acceleration to beat China).
The bill faces a difficult procedural path in a Republican-controlled Congress that has already struggled to unify around far narrower AI measures. Reporting identifies competing proposals already in play: Sens. Josh Hawley (R-Mo.) and Richard Blumenthal (D-Conn.) have a bill directing the Department of Energy to evaluate frontier models; Sen. Mark Warner (D-Va.) has legislation to establish an AI risk board; and Hawley separately has pushed the GUARD Act targeting AI companion chatbots and children. Congress has also recently fought over a competing federal preemption question — a 10-year moratorium on state-level AI regulation that passed the House Energy and Commerce Committee and cleared the House by a narrow 215-214 vote in a reconciliation vehicle in May 2026, before facing bipartisan Senate pushback. That preemption fight and the Sanders-Casar ban represent opposite regulatory instincts (blocking state action vs. imposing the most restrictive possible federal action) within the same Congress.
Politically, the bill is widely read as a marker for the progressive wing of the Democratic Party heading into the November 2026 midterms and the 2028 presidential cycle, rather than a bill with a plausible path to floor votes this Congress. Casar chairs the Congressional Progressive Caucus. Multiple 2028-aligned Democratic governors — reporting names Colorado's, Pennsylvania's, and Virginia's governors as having taken AI-oversight-related action or made statements — have moved on AI in recent weeks, suggesting the issue is becoming a defining marker within the party ahead of the primary contest. Republican reaction has ranged from dismissive (Sen. Rick Scott calling the Senate incapable of passing anything) to partially sympathetic on substance but opposed to a ban (Sen. Josh Hawley, who backs liability-based approaches instead).
Intersection Groups (7)
Proximity: DirectNear-TermFLOW C
Sen. Bernie Sanders / Rep. Greg Casar (bill sponsors)
Sanders and Casar have staked their political capital on the most restrictive AI proposal in Congress, pairing a permanent superintelligence ban with 20-year prison terms and a corporate death penalty modeled on nuclear-weapons law. As chair of the Congressional Progressive Caucus, Casar's involvement signals the bill functions as a coalition-building vehicle within House progressives rather than a bill designed to attract Republican cosponsors.
Strategic Options
01Narrow the bill's definition of artificial superintelligence in a revised draft to address the ambiguity critics have already raised, following the pattern of iterative revision Hawley/Blumenthal used narrowing DOE frontier-model evaluation scope.
02Pursue a discharge petition or attach provisions to a must-pass vehicle only if bipartisan cosponsors emerge from the Hawley-aligned liability camp rather than seeking standalone floor time.
03Use the bill as a 2026 midterm and 2028 primary positioning document, coordinating messaging with Democratic governors who have already taken AI-oversight action.
↳ [ASSESSED] The bill functions less as legislation with a floor pathway than as a definitional marker forcing every other AI proposal in Congress — Hawley-Blumenthal's DOE evaluation bill, Warner's risk board, Hawley's GUARD Act — to be read against the most extreme option now on the table.
FLOW Rationale: Sponsors face a genuinely unresolved question of how to operationalize a 'superintelligence' definition industry critics call unworkably broad, and the bill's fate depends on interconnected dynamics across committee assignment, coalition-building with skeptical Republicans, and the 2028 primary calendar — all without an established playbook.
Scale (Moderate): The bill reframes the AI-regulation debate around prohibition rather than disclosure or liability, but as a single-sponsor progressive bill it lacks committee traction to become binding policy this Congress.
Complexity (High): Sponsors must navigate an unsettled definition of 'superintelligence' that critics say could sweep in far less autonomous systems, undermining the bill's own enforceability.
Key Question
Can Sen. Bernie Sanders and Rep. Greg Casar narrow the Ban Artificial Superintelligence Act's definition of artificial superintelligence enough to answer critics like AI researcher Gary Marcus without weakening the bill's core prohibition?
Watch Signals:- [Possible] Introduction of a revised bill text narrowing the superintelligence definition — no committee referral action has been reported as of September 23, 2026, so timing is unanchored.
- [Possible] Additional House or Senate cosponsors signing on beyond Sanders and Casar; current reporting names no cosponsors beyond the two lead sponsors.
- [Unlikely] A committee hearing scheduled specifically on this bill before the November 2026 midterms, given reported House floor-time constraints through the election recess.
Proximity: DirectNear-TermFLOW D
The White House
The administration's newly announced but structurally undefined 'AI Force' and forthcoming AI czar directly compete with the bill's proposed Department of Artificial Intelligence for the same regulatory turf, forcing the administration to either ignore the bill publicly or define its own initiative's authority in response. The President has already called superintelligence fears a 'hoax' and a 'SICK conspiracy,' setting up an unusually direct rhetorical collision with sponsors who compare the risk to nuclear proliferation.
Strategic Options
01Name the AI czar promptly and publicly define the AI Force's authorities to pre-empt comparisons to the bill's proposed Department of Artificial Intelligence, following the precedent of naming David Sacks quickly as the prior White House AI and Crypto Czar.
02Direct agencies to rely on existing criminal and civil enforcement authority as the administration has stated, rather than seeking new statutory authority that would require engaging with the Sanders-Casar bill's framework.
03Coordinate with House Speaker Mike Johnson's planned meeting with AI industry leaders to present a unified anti-regulation front ahead of the midterms.
↳ [ASSESSED] The administration's 'AI Force' announcement, which provided no detail on structure or authority, was made days before the Sanders-Casar bill's formal unveiling — meaning the administration's own initiative is being defined reactively in the shadow of the more detailed legislative proposal it opposes.
FLOW Rationale: The scale is large because the collision touches the administration's entire AI competitiveness strategy against China, and the complexity is high because the AI Force has no defined authorities yet, forcing improvised policy design under public pressure.
Scale (Large): The dispute reaches into the executive branch's own AI governance architecture and national competitiveness posture relative to China, not a single program.
Complexity (High): The administration has provided no details on the AI Force's authorities, budget, or czar's responsibilities, meaning any response to the Sanders bill must be improvised without an existing structural counterpart.
Key Question
What legal authority will the AI czar and AI Force possess once named, given the administration's stated reliance on existing criminal and civil law rather than new statutory power?
Watch Signals:- [Likely] Naming of a new AI czar — the President stated on September 19-20, 2026 this would happen 'in the near future,' making an announcement within weeks plausible based on that explicit timeline.
- [Possible] Formal structure or executive order establishing the AI Force's authorities and budget, which has not yet been detailed as of September 23, 2026.
- [Possible] House Speaker Mike Johnson's reported meeting between the President and AI industry leaders, which Johnson said was agreed to happen 'within the next week' as of a report dated within the past two weeks.
Proximity: CloseMonitorFLOW B
House Republican leadership (Speaker Mike Johnson)
Johnson has publicly stated the President is not interested in new federal AI regulation and that Congress 'cannot have a moratorium on the development of AI,' placing House Republican leadership on record against the bill's core premise before it even reaches committee. Johnson's agreement to arrange a meeting between the President and AI industry leaders signals leadership is managing this as an industry-relations issue rather than a legislative one.
Strategic Options
01Continue declining to schedule hearings or markups on the bill, consistent with the pattern set when the House majority narrowly muscled through the competing state-AI-moratorium provision in May 2026 rather than engaging restrictive proposals directly.
02Use the scheduled meeting with AI industry leaders to publicly counter-message against the bill's prohibition framework before the midterms.
03Redirect committee attention toward narrower bipartisan proposals such as the Hawley-Blumenthal DOE frontier-model evaluation bill to demonstrate action without embracing a ban.
↳ [ASSESSED] Leadership's anti-moratorium rhetoric this month directly echoes the position taken during the House's narrow 215-214 vote in May 2026 to advance a 10-year state AI moratorium, showing a consistent institutional preference for less restriction rather than a case-by-case reaction to this bill specifically.
FLOW Rationale: Leadership's opposition is a straightforward continuation of an already-established position, giving House Republicans a clear, low-complexity path of simply not advancing the bill through committee.
Scale (Moderate): Leadership's public opposition effectively forecloses floor time for the bill this Congress without requiring a formal committee vote.
Complexity (Low): Leadership's position is already publicly stated and consistent with the party's established anti-moratorium stance, requiring no new procedural navigation.
Key Question
Will House Speaker Mike Johnson allow any committee hearing on the Ban Artificial Superintelligence Act before the November 2026 midterm recess, given his stated position against an AI development moratorium?
Watch Signals:- [Unlikely] A House Energy and Commerce Committee hearing scheduled on the Sanders-Casar bill before the midterm recess, given leadership's stated opposition to any moratorium framework.
- [Likely] The reported Trump-AI industry leaders meeting occurring within the near-term window Johnson described, given it was already agreed to as of a report within the past two weeks.
Proximity: DirectNear-TermFLOW D
Frontier AI developers (OpenAI, Anthropic, Google DeepMind)
The bill's 'corporate death penalty' provision would strip a violating firm's legal authority to operate in the United States, and its temporary pause on advanced AI development pending new federal safety rules would directly halt frontier model development at the named labs if enacted. Notably, the bill's sponsors cite endorsements from employees at leading AI companies and reference a former Anthropic researcher's warnings, showing the bill draws on internal industry dissent even as company leadership publicly resists prohibition-style regulation.
Strategic Options
01Commission independent legal analysis of whether current frontier models would trigger the bill's broad superintelligence definition, mirroring how AI researcher Gary Marcus has already publicly flagged the definitional ambiguity as the bill's central flaw.
02Engage the Hawley-Blumenthal DOE frontier-model evaluation framework as a preferred alternative regulatory vehicle to pre-empt the more restrictive Sanders-Casar approach.
03Coordinate public messaging distancing company leadership from internal employee endorsements of the bill to avoid signaling institutional support for a prohibition framework.
↳ [CONFIRMED] Multiple employees at leading AI companies are endorsing the bill even as their own executives publicly resist prohibition-based regulation, revealing an internal industry split that a purely company-level analysis of AI policy positioning would miss.
FLOW Rationale: The scale is large because the bill targets the core commercial activity of the frontier AI industry outright, and complexity is high because no legal precedent exists for applying a 'matching or exceeding human cognitive performance' threshold to real products.
Scale (Large): A prohibition and pause targeting the frontier AI industry's core product development would, if enacted, halt the primary commercial activity of the largest AI labs.
Complexity (High): Labs face an unresolved definitional question — where matching-or-exceeding-human-cognition thresholds apply — that determines whether current or near-term models fall under the ban, with no legal precedent to guide compliance planning.
Key Question
Would current frontier models from OpenAI, Anthropic, or Google DeepMind meet the Ban Artificial Superintelligence Act's definition of AI capable of matching or exceeding human cognitive performance across a wide range of tasks?
Watch Signals:- [Possible] Public statements from frontier lab executives responding directly to the bill's corporate death penalty provision, following the pattern of Anthropic's CEO already telling media there are 'real dangers' with AI while stopping short of endorsing a ban.
- [Possible] Additional current or former employees at frontier labs publicly endorsing the bill, following the pattern already reported of multiple employees signing on.
Proximity: CloseMonitorFLOW B
Sen. Josh Hawley (R-Mo.)
Hawley has publicly said the underlying concerns about superintelligence are warranted but stopped short of backing a ban, instead proposing that AI companies be held liable for harms — positioning him as the pivotal Republican voice who could either isolate the Sanders-Casar bill as too extreme or lend it partial legitimacy by acknowledging shared concerns. Hawley separately has his own DOE frontier-model evaluation bill with Sen. Richard Blumenthal and the child-safety-focused GUARD Act, giving him a competing legislative vehicle to steer AI-safety momentum away from prohibition.
Strategic Options
01Advance the Hawley-Blumenthal DOE frontier-model evaluation bill as the moderate alternative to the Sanders-Casar prohibition, capitalizing on Hawley's existing bipartisan cosponsorship with Blumenthal.
02Publicly distinguish liability-based accountability from outright prohibition in floor remarks to prevent the Sanders-Casar bill from defining the Republican AI-safety conversation.
03Use the GUARD Act's child-safety framing as a lower-controversy vehicle to demonstrate Republican willingness to regulate AI without embracing a development ban.
↳ [CONFIRMED] Hawley's acknowledgment that superintelligence concerns are warranted, paired with his refusal to back a ban, positions him as the likely broker of any eventual bipartisan AI-safety compromise that emerges once the Sanders-Casar bill's prohibition approach is set aside.
FLOW Rationale: Hawley's position is moderate in scale because it shapes the competing bipartisan alternative rather than binding law, and low complexity because his liability-based stance is already clearly articulated and requires no new procedural navigation.
Scale (Moderate): Hawley's position shapes which bipartisan AI-safety vehicle gains traction but does not itself carry binding legal force.
Complexity (Low): Hawley has already staked out a clear liability-based alternative position, giving him an established path to differentiate from the Sanders-Casar approach without new procedural complications.
Key Question
Will Sen. Josh Hawley's DOE frontier-model evaluation bill with Sen. Richard Blumenthal gain committee traction as the moderate alternative to Sen. Bernie Sanders' Ban Artificial Superintelligence Act?
Watch Signals:- [Possible] A Senate Commerce or Energy Committee hearing scheduled on the Hawley-Blumenthal DOE frontier-model evaluation bill, positioning it against the Sanders-Casar bill for bipartisan momentum.
- [Possible] Additional Republican cosponsors joining Hawley's liability-based approach as a counter to the prohibition framework.
Proximity: AffectedMonitorFLOW B
Information Technology and Innovation Foundation (ITIF) and industry policy critics
ITIF President Daniel Castro has already publicly warned that a unilateral U.S. ban would not be matched by China, creating an economic and national security cost from ceding AI leadership — a direct rebuttal to the bill's international-agreement provisions that assume other nations would follow a U.S. prohibition. This positions industry-aligned think tanks as an organized voice actively working to frame the bill as a competitiveness risk ahead of any committee consideration.
Strategic Options
01Publish formal comparative analysis of U.S. versus Chinese AI investment levels to reinforce the competitiveness argument against the bill ahead of any hearing.
02Coordinate with Republican committee staff to supply competitiveness-focused testimony if the bill ever receives a hearing.
03Target messaging at swing Democratic senators who may be wary of a China-competitiveness argument ahead of the midterms.
↳ [SINGLE-SOURCE] ITIF's competitiveness critique, delivered directly in response to Sanders' proposal, previews the likely core line of attack industry-aligned groups will use in any future committee hearing, regardless of the bill's odds of advancing.
FLOW Rationale: ITIF's role is moderate in scale, shaping debate framing rather than legislative text, and low complexity because the competitiveness critique is an already-established policy argument requiring no new analytical framework.
Scale (Moderate): ITIF's advocacy shapes the competitiveness framing used against the bill in Congress but does not itself alter legislative outcomes directly.
Complexity (Low): The China-competitiveness argument is an established critique requiring no novel policy analysis to deploy against the bill.
Key Question
Does the Ban Artificial Superintelligence Act's international-agreement provision include any enforcement mechanism to ensure China or other AI powers would match a U.S. prohibition, as ITIF President Daniel Castro's competitiveness critique implies it currently lacks?
Watch Signals:- [Possible] Formal ITIF or similar think-tank publications quantifying the competitiveness cost of a unilateral U.S. AI development pause.
- [Possible] Republican committee staff citing ITIF's competitiveness argument in any future hearing materials on the bill.
Proximity: AffectedMonitorFLOW C
Democratic governors with 2028 ambitions
Reporting identifies Democratic governors who have taken AI-oversight action in the weeks preceding the bill's unveiling, including one who signed an executive order urging consideration of an AI 'kill switch' and another who has called on the federal government to regulate AI — positioning state-level executive action as a parallel track to the Sanders-Casar federal bill within the same intra-party 2028 competition.
Strategic Options
01Issue further state-level executive actions on AI oversight that stop short of the Sanders-Casar bill's outright prohibition, positioning as safety-conscious without adopting the more politically risky ban framework.
02Publicly respond to the Sanders-Casar bill to define a distinct 2028 AI-policy lane rather than being grouped with either the prohibition or anti-regulation poles.
03Coordinate messaging with Congressional Progressive Caucus Chair Rep. Greg Casar to align state and federal progressive AI positioning ahead of primary season.
↳ [ASSESSED] The convergence of multiple Democratic governors taking AI-oversight action in the same week as the Sanders-Casar bill's unveiling suggests AI policy is becoming a defining pre-primary marker for 2028 contenders rather than a niche issue.
FLOW Rationale: The scale is moderate because governors' actions are state-level rather than federal, but complexity is high because each must navigate 2028 primary positioning against two opposing federal poles without an established playbook for how AI ranks against other primary issues.
Scale (Moderate): Governors' actions operate at the state executive level and shape 2028 primary positioning rather than altering federal law directly.
Complexity (High): Each governor must calibrate AI-oversight positioning against both the Sanders-Casar federal prohibition and the administration's competing anti-regulation posture without a clear precedent for how the issue will play in a Democratic primary.
Key Question
Which Democratic governors positioning for the 2028 presidential primary will publicly endorse or distance themselves from Sen. Bernie Sanders and Rep. Greg Casar's Ban Artificial Superintelligence Act?
Watch Signals:- [Possible] Additional Democratic governors issuing AI-oversight executive orders or public statements in the weeks following the bill's September 23, 2026 unveiling.
- [Possible] Direct public statements from Democratic governors specifically addressing the Sanders-Casar bill's prohibition framework rather than general AI oversight.
The claims behind this analysis, each with its verification status — including what is contested, unverified, or could not be established.
What each grade meansSen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) formally unveiled the Ban Artificial Superintelligence Act on September 23, 2026, with the bill text provided first to the Associated Press.
Establishes the exact bill and date anchoring the entire event and every downstream intersection.
The bill sets penalties of a 'corporate death penalty' for violating entities and up to 20 years in prison for individuals, described by the sponsors' own release as similar to existing penalties for unlawfully developing nuclear weapons.
Defines the severity of the enforcement regime that drives the Large-scale classification for frontier AI developers and the White House.
President Trump announced on September 19-20, 2026 that he would form an 'AI Force' modeled on the Space Force and name a new AI czar, while calling superintelligence fears a 'hoax' and vowing not to 'hinder or stifle' AI industry growth.
Establishes the direct competing policy posture from the executive branch that structures the White House and House leadership intersections.
A proposed 10-year moratorium on state-level AI regulation passed the House by a narrow 215-214 vote on May 22, 2026 as part of a reconciliation bill, before facing bipartisan pushback in the Senate.
Shows Congress has already fought a closely-divided battle over the opposite regulatory instinct (blocking state action) in the same session, contextualizing how unlikely the more restrictive Sanders-Casar bill is to advance.
Sens. Josh Hawley (R-Mo.) and Richard Blumenthal (D-Conn.) have introduced a competing bill that would direct the Department of Energy to establish an evaluation program reviewing frontier AI models, modeled on FDA-style review.
Identifies the moderate bipartisan alternative vehicle that could absorb AI-safety momentum instead of the Sanders-Casar prohibition, anchoring the Hawley intersection.
House Speaker Mike Johnson stated Congress 'cannot have a moratorium on the development of AI' and said the President agreed to meet with AI industry leaders 'within the next week,' as reported in coverage from within the past two weeks.
Establishes House Republican leadership's explicit opposition to the bill's core premise, driving the low-complexity classification for that intersection.