Biosecurity monitoring systems can aggregate separate risk signals across proprietary AI query logs, mail-order DNA synthesis orders, and outsourced cloud laboratory requests

Overview Debates Related claims Topics

Fine-tuning open AI models on small, targeted biological datasets can restore dangerous dual-use capabilities excluded during initial pre-training

Joshua Monrad categorizes biological risks into four phases: human-created bioweapons, AI-enabled human uplift, AI agents socially engineering humans, and end-to-end autonomous AI bioweapon execution

Two individuals who detonated a chemical explosive device at a California fertility clinic utilized a large language model to inform their attack

Sentinel Bio has raised tens of millions of dollars and issued approximately 100 philanthropic grants for biosecurity projects as of October 2026

Controlled pre-training experiments on biological foundation models ESM and Evo demonstrated that withholding specific virus datasets limited human virology capabilities while preserving broader biological functionality

Joshua Monrad dates the emergence of the phase where humans use AI systems for biological weapon design assistance to the late 2010s and early 2020s

A biological uplift study co-funded by Sentinel Bio and the Frontier Model Forum and conducted by ActiveSight evaluated 150 participants in a physical wet lab at a cost exceeding one million dollars

A biosecurity evaluation conducted by Anthropic and Deloitte tested the Mythos model on virology protocol execution using a sample size of fewer than 30 PhD-level participants

As of October 2026, the development of end-to-end autonomous biological weapons without humans in the loop is constrained by robotics capabilities rather than cognitive capabilities

Joshua Monrad supports enacting a temporary pause on frontier AI biological model training until comprehensive societal biodefenses and supply chain safeguards are operational

Joshua Monrad and Sentinel Bio conducted a systematic evaluation of biosecurity physical choke points that included cell lines, mammalian animal models, pipettes, and biosafety cabinets

The International Mission Board's AI charter prohibits giving artificial intelligence autonomous authority in high-risk and high-stakes contexts

The International Mission Board requires human-in-the-loop oversight whenever artificial intelligence is used to generate code

The International Mission Board's AI policy prohibits artificial intelligence from making final autonomous decisions about people

The International Mission Board's AI charter prohibits artificial intelligence from misrepresenting itself as human in ministry

California Senate Bill 1047 included a mandatory kill switch liability for artificial intelligence developers

In an analysis of approximately 3,600 messages from AI power user group chats, 22% of messages identified trust as their primary concern regarding AI agents

Technology entrepreneur Shane Mac introduced an agent containment concept called Picnic, comparing agent boundary setting to fencing a yard for children

Anthropic defines its recursive self-improvement danger threshold as achieving as much AI capability progress in one year as occurred between 2018 and 2024

Anthropic updated its responsible scaling policy disclosures to state Anthropic cannot safely use Claude to fully autonomously design and build frontier models without meaningful human supervision