What happened
Senator Bernie Sanders and Representative Greg Casar have introduced the 'Ban Artificial Superintelligence Act,' a legislative proposal that seeks to prohibit the development of artificial superintelligence and impose a moratorium on advanced AI systems. The bill, which was shared with the Associated Press, proposes the creation of a new federal Department of Artificial Intelligence to oversee compliance and establish regulatory frameworks.
The proposed legislation defines artificial superintelligence as any system that exceeds human cognitive performance in most domains or possesses the capability to plan the destruction or disempowerment of humanity. It also mandates an immediate pause on the development of advanced AI systems, defined as those trained on 10^25 or more operations, until a new Department of Artificial Intelligence is fully staffed and has established regulatory rules.
Under the bill, companies that violate the ban or the pause could face severe penalties, including up to 20 years in prison for individuals and the forfeiture of intellectual property and assets for corporations. The bill also includes whistleblower protections for employees who report risks associated with superintelligence to the new department.
The legislation advocates for international agreements and export controls on AI computing hardware to prevent the development of superintelligence globally. Senator Sanders has explicitly compared the need for international cooperation on AI to Cold War-era nuclear arms control.
The bill has received support from individual employees at major AI firms, including researchers from OpenAI and Google DeepMind, who provided statements to the Associated Press. However, the bill faces long odds in Congress, with three other AI-related bills currently awaiting action without scheduled votes.
Source details: thenextweb.com
Why it matters
This legislation represents a significant escalation in the legislative debate over , moving beyond standard oversight to propose criminal penalties and the seizure of corporate assets. By defining 'advanced AI' through specific computational thresholds (10^25 operations) and targeting 'recursive self-improvement,' the bill attempts to codify technical boundaries for AI development. The proposal highlights a growing divide in US policy, contrasting sharply with the current administration's stated goal of accelerating AI development to maintain global competitiveness. The bill's inclusion of a 'corporate death penalty'—whereby companies violating the ban would forfeit their intellectual property and assets to the federal government—marks a radical shift in how lawmakers are approaching the governance of private technology firms. Furthermore, the bill's focus on international export controls and cooperation with nations like China underscores the geopolitical stakes involved in AI development, framing the technology as a potential existential risk comparable to nuclear weapons.
The bill introduces a 'corporate death penalty' for non-compliance, which would involve the federal government seizing a company's assets and intellectual property. This is a departure from traditional regulatory fines and signals a move toward treating AI development as a matter of national security.
The proposal creates a direct policy conflict with the current administration. President Trump recently stated at the UN that the US intends to encourage AI development, which he referred to as 'super intelligence,' rather than restricting it. This creates a clear political divide regarding the future of US AI strategy.
The bill's reliance on a specific computational threshold (10^25 operations) to define 'advanced AI' provides a concrete, albeit controversial, metric for regulation. This approach attempts to move the conversation from abstract safety concerns to measurable technical limits, though it faces criticism regarding the potential for US companies to fall behind international competitors.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
What to watch next
The bill faces significant legislative hurdles in the Republican-controlled Congress, where other AI-related proposals have already stalled. Observers should monitor whether this bill gains traction as a platform for 2028 presidential candidates, as suggested by Senator Sanders, or if it remains a symbolic effort. Additionally, the industry response from major AI labs—particularly regarding the proposed 10^25 operations threshold—will be critical, as this metric would effectively halt the training of current frontier models. The potential for conflict between this legislative push and the administration's stated policy of encouraging AI growth remains a primary point of friction.
Watch for the bill's reception in the House and Senate committees. Given the current legislative climate and the lack of movement on existing AI bills, the likelihood of this legislation passing is low, but it may serve as a for future policy debates.
Monitor the reaction of the AI industry, particularly regarding the feasibility of the 10^25 operations limit and the potential impact on research and development cycles.
Observe how the 2028 presidential candidates address this bill, as Senator Sanders has indicated he intends to make AI regulation a central issue for the upcoming election cycle.