Follow us on Google News
A renewed push to pause the development of advanced artificial intelligence has emerged in the United States after Democratic gains in the midterm elections, with lawmakers calling for stricter controls on frontier AI systems and a prohibition on artificial superintelligence.
Congresswoman Yassamin Ansari said on November 4 that a pause on advanced AI development, stronger regulation of the industry and a ban on artificial superintelligence should be part of the Democratic agenda. Her remarks revived attention around legislation introduced earlier this year that would impose sweeping restrictions on the development and deployment of highly capable AI systems.
However, no such ban has been enacted. The proposals remain bills before Congress and would need to pass through the legislative process before becoming law.
Bill Would Pause Advanced AI Development
The House proposal, H.R. 10538, known as the Ban Artificial Superintelligence Act of 2026, was introduced on September 24 by Representative Greg Casar, with Ansari among its co-sponsors. A companion Senate bill, S. 5493, was introduced by Senator Bernie Sanders on September 23.
The legislation would create a new federal Department of Artificial Intelligence and impose a mandatory pause on systems classified as advanced AI until the department is fully established and safety regulations are in place.
Under the bill, an advanced AI system is defined primarily by the amount of computing power used during training. The threshold is set at at least 10²⁵ integer or floating-point operations, with provisions allowing the threshold to be adjusted over time as training methods become more efficient.
During the proposed pause, advanced AI systems could not be trained, modified or fine-tuned, except for limited purposes such as removing capabilities considered precursors to superintelligence.
The legislation would also prevent unreleased advanced systems from being deployed until the proposed AI department establishes rules governing safe development, monitoring, audits and pre-deployment approval.
Superintelligence Would Face an Outright Ban
The legislation goes beyond temporarily slowing frontier AI development.
It would prohibit people from developing, deploying, acquiring, possessing, funding, importing or transferring artificial superintelligence or AI systems displaying characteristics that lawmakers believe could lead to superintelligence.
The bill defines artificial superintelligence as an AI system that exceeds human cognitive performance across most areas or possesses capabilities sufficient to plan and execute the destruction or disempowerment of humanity.
It also identifies several so-called precursor characteristics, including the ability to accelerate AI research, gain unauthorised access to infrastructure, resist attempts to shut it down, enhance weapons development, modify its own functions or evade human oversight.
Under the proposal, developers of advanced AI models would eventually need a federal charter and could not deploy an advanced system without government approval.
Violations Could Bring Severe Penalties
The proposed enforcement provisions are unusually strict.
Covered policymakers or so-called rogue actors who recklessly violate key parts of the legislation could face fines and imprisonment for up to 20 years.
Individuals outside policymaking roles who recklessly violate the relevant provisions could be barred from working in the AI industry for 10 years, subject to appeal and limited exceptions.
The legislation would also give the proposed Department of Artificial Intelligence significant oversight powers, including the ability to inspect advanced AI systems and require pre-deployment approval.
If a system were identified as displaying prohibited precursor characteristics, the department could order it isolated from the internet and potentially rendered inoperative if the risks could not be removed.
Debate Over AI Risk Remains Divided
The push comes amid continuing disagreement over what poses the greatest danger from advanced AI.
Some technology executives and researchers have warned that increasingly capable systems could eventually exceed human control, while others argue that the more immediate threats come from human misuse, corporate concentration, misinformation, cybercrime and poor governance.
The proposed legislation takes the more restrictive approach by treating the development of superintelligence itself as a potential public safety threat.
It also reflects growing political pressure for governments to intervene before frontier AI systems become significantly more capable.
Bill Has Not Yet Become Law
Despite renewed attention, the proposals remain at an early legislative stage.
H.R. 10538 was referred to the House Committee on Oversight and Government Reform, while S. 5493 was referred to the Senate Committee on Commerce, Science and Transportation. Neither bill has passed its respective chamber.
That means current AI models are not facing an immediate federal development ban under these bills.
The significance of Ansari’s latest remarks lies instead in the possibility that proposals once viewed as aggressive restrictions on frontier AI could receive greater political attention following changes in congressional control.
Whether lawmakers ultimately support a mandatory pause, a permanent prohibition on superintelligence or a less restrictive regulatory framework will depend on how the legislation develops in Congress.
For now, the debate marks one of the clearest efforts yet by US lawmakers to place hard legal limits on the development of the most powerful AI systems.
About the author — Suvedita Nath is a science student with a growing interest in cybercrime and digital safety. She writes on online activity, cyber threats, and technology-driven risks. Her work focuses on clarity, accuracy, and public awareness.
