U.S. Bill Proposes Permanent Ban on Artificial Superintelligence and Creation of New Federal AI Agency
A new bipartisan bill in the U.S. Senate would permanently ban the development and deployment of artificial superintelligence (ASI) and create a new federal agency to oversee all advanced AI systems. Introduced by Senator Michael Bennet (D-CO) and co-sponsored by Senator Gary Peters (D-MI), the legislation defines ASI as any AI that can outperform humans in every economically valuable task. The bill immediately establishes a Federal Artificial Intelligence Agency (FAIA) to enforce the ban and regulate high-risk AI.
“We cannot gamble with the future of humanity by allowing unaccountable corporations to build intelligence that exceeds our own,” Senator Bennet said in a statement.
The Bill’s Key Provisions
The legislation imposes a strict, permanent prohibition on ASI. It also creates a licensing and oversight framework for all advanced AI models.
- Permanent ban on ASI – No entity may develop, train, or deploy an AI system that meets the legal definition of superintelligence. Violators face heavy fines and potential criminal penalties.
- New Federal Artificial Intelligence Agency – The FAIA will be an independent regulatory body with authority to test, license, and audit AI systems before they reach the market.
- Mandatory safety testing – Companies must submit high-risk AI models for pre-market review. Testing covers capability thresholds, alignment, and potential for autonomous replication.
- Transparency and accountability rules – Developers must disclose training data sources, compute usage, and model architecture. A public registry will track all licensed systems.
- Prevention of monopolistic control – The bill includes provisions to break up concentrated ownership of AI infrastructure, including compute clusters and training datasets.
What Is Artificial Superintelligence?
The bill defines ASI as “any artificial intelligence system that demonstrably exceeds human cognitive performance across all economically valuable occupations and tasks.” This broad definition aims to capture future systems, not just current ones.
The authors argue that ASI represents an existential threat because it could rapidly improve itself and outmaneuver human oversight. The ban is intended as a permanent safeguard, not a temporary pause.
New Federal AI Agency
The Federal Artificial Intelligence Agency would be modeled after the Food and Drug Administration and the Federal Aviation Administration. It would have three core divisions:
- Capability Assessment – Evaluates whether a given AI system approaches or exceeds superintelligence thresholds.
- Safety and Alignment – Tests whether systems act in accordance with human values and can be reliably controlled.
- Licensing and Compliance – Issues operating licenses, conducts audits, and enforces penalties for violations.
The agency would be funded through a fee on AI companies proportional to their compute usage and model size.
Support and Opposition
Supporters say the bill is necessary to prevent a technological race toward uncontrollable AI. Several AI safety researchers and civil society groups have endorsed the proposal.
Opponents argue that a permanent ban stifles innovation and risks driving development overseas. Some technology companies warn that the definition of ASI is too vague and could capture systems that are merely advanced but not dangerous.
The bill faces uncertain odds in Congress. It has been referred to the Senate Commerce Committee, where heated debate is expected.
What Happens Next
The bill enters committee hearings this fall. If passed, the FAIA would need to be operational within two years. The permanent ban on ASI would take effect immediately upon enactment.
Debate over the definition of superintelligence and the agency’s enforcement powers is likely to dominate the legislative process.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.