# Former White House Senior Director Calls for Mandatory AI Triggers

By Riza Dagoc

Canonical URL: https://www.tokenpost.com/news/regulation/28909
Published: 2026-10-09T11:49:41.000Z
Updated: 2026-10-09T11:49:41.000Z
Section: Regulation

> Raj Panjabi says AI-biosecurity safeguards should link observable warning signs to predetermined actions as systems lower barriers to dangerous biological work.

Raj Panjabi, the former White House senior director for global health security and biodefense, is calling for AI-biosecurity rules that connect observable warning signs to mandatory actions as artificial intelligence lowers barriers to dangerous biological work.

Panjabi said the key risk is not only that AI can answer sensitive biological questions. It is that people without advanced expertise may use the technology to overcome laboratory barriers and move digital biological designs into the physical world.

“Not yet,” Panjabi said when discussing the current lack of efficient human-to-human H5N1 transmission. “Not yet is useful information. It is not a plan.”

His proposed framework would pair specific signals with predetermined responses. Examples include nonexperts using AI to solve experimental problems, AI-generated biological sequences being ordered from synthesis providers or used in automated laboratories, and evidence that AI has reduced the practical experience needed for dangerous biological work.

The physical-world bottleneck remains important because digital designs still require materials, equipment and laboratory facilities. The Bipartisan Commission on Biodefense lists AI as a technology priority and recommends model evaluations, safeguards, disease surveillance, microbial forensics and adaptive nucleic-acid screening.

The commission has also warned that AI systems can compress timelines for biological work that previously required specialized expertise and costly infrastructure.

Anthropic’s September 2026 threat report describes five biological-misuse cases involving its models between December 2025 and August 2026. One involved a researcher planning avian-influenza mammalian-adaptation experiments with Claude. Anthropic said its classifiers restricted the activity to weaker models.

A bipartisan Senate bill introduced Feb. 4 would require gene-synthesis providers to screen customers and orders for dangerous pathogens. Jeff Alstott’s Senate testimony recommended know-your-customer requirements for gene-synthesis providers, cloud laboratories and benchtop synthesizers, along with screening synthesized genetic material above a specified threshold.

Panjabi’s approach would extend AI safeguards beyond model behavior to DNA-synthesis screening, laboratory automation and other controls governing access to physical biological capabilities. The framework is intended to establish the response before a biological threat becomes a crisis.