AI Standards has announced a report by Chief Vision Officer Fran Horvath arguing that verifiable technical controls offer a more workable path for advanced AI oversight than a legislative pause.
-- AI Standards has announced the publication of a new policy and engineering report examining proposals in Congress to pause advanced artificial intelligence development. Written by Chief Vision Officer Fran Horvath, the report argues that a domestic moratorium would be difficult to enforce against actors outside U.S. jurisdiction, and proposes cryptographic and runtime verification controls as an alternative approach to AI oversight.
The full report is available at https://aistandards.blog/why-you-cannot-pause-advanced-ai-the-physics-of-contagion-global-diffusion-and-runtime-proof
The report arrives as AI regulation moves up the agenda following the November 2026 midterm elections. An NBC News exit poll found that 56% of voters described AI and data centers as a crucial or major factor in their vote for Congress. Shortly after the election, Rep. Yassamin Ansari called a pause on advanced AI development a day-one priority for the incoming Democratic majority.
That call centers on the Ban Artificial Superintelligence Act of 2026, introduced by Sen. Bernie Sanders as S. 5493 on September 23 and by Rep. Greg Casar as H.R. 10538 the following day. According to the bill text, the legislation would create a cabinet-level Department of Artificial Intelligence and pause development of covered advanced AI systems, defined by a training threshold of 10^25 operations, until the new department establishes safety rules. Supporters say the measure is needed to protect the public from the risks of increasingly capable systems. Both bills remain in committee.
Horvath’s report takes a different view. “Attempting to halt advanced AI with a legislative pause is akin to tossing a handful of viruses into a closet, shutting the door, and thinking you have controlled a contagion,” Horvath writes. The report contends that because AI capabilities spread through software and open-weight models across borders, a pause would apply mainly to developers who already follow the rules.
To support that argument, the report cites recent third-party research. In October 2026, Epoch AI reported that GLM 5.3, an open-weight model released by Z.ai, built working exploits against six Firefox releases using security fixes that had been published but not yet shipped to users, at a marginal cost of about $1,000 per release. Wiz Research documented security incidents at nine AI developer platforms that month, including a case in which a single stolen cloud key generated about 1.7 million AI model calls in 13 hours. Google Threat Intelligence Group reported that attackers using the open-source agentic framework ARTEX typically moved from a first compromised account to deep system access in around four hours, and in one case in 15 minutes.
The report also questions compute thresholds as a regulatory tool, noting that many current cyber operations rely on smaller models running agent-based workflows, which need far less computing power than the large training runs such thresholds are designed to capture.
As an alternative, the report outlines three areas of technical control. The first is hardware-rooted cryptographic identity, linking AI agents and model weights to verifiable credentials and accountable parties. The second is continuous runtime monitoring designed to detect behavioral drift and covert data leakage in deployed systems. The third is zero-knowledge governance records, in which authorizations and guardrail checks are recorded in tamper-evident, ledger-anchored audit trails that regulators could verify without accessing private enterprise data.
About AI Standards
AI Standards is a Delaware-registered company developing AI governance infrastructure and private AI deployments for enterprises. The company was co-founded by Chief Executive Officer Joe Quenneville, Chief Vision Officer Fran Horvath, and Chief Technology Officer and Chief Scientist Anwesh Rath. Its work in cryptographic audit trails, model integrity verification, and runtime monitoring is covered by more than 115 patent-pending applications across more than 35 technology domains.
More information can be found at https://aistandards.blog/why-you-cannot-pause-advanced-ai-the-physics-of-contagion-global-diffusion-and-runtime-proof
Contact Info:
Name: Fran Horvath
Email: Send Email
Organization: AI Standards Inc
Address: 311 13th Ave South, Lanett, Alabama 36863-2220, United States
Phone: +1-678-205-9720
Website: https://aistandards.inc
Source: PressCable
Release ID: 89209250
In case of identifying any errors, concerns, or inconsistencies within the content shared in this press release that necessitate action or if you require assistance with a press release takedown, we strongly urge you to notify us promptly by contacting [email protected] (it is important to note that this email is the authorized channel for such matters, sending multiple emails to multiple addresses does not necessarily help expedite your request). Our expert team is committed to addressing your concerns within 8 hours by taking necessary actions diligently to rectify any identified issues or supporting you with the removal process. Delivering accurate and reliable information remains our top priority.
