KYC Becomes KYA: Sentinel Bio's Joshua Monrad on DNA Screening, Uplift Studies & Agent Monitoring

In a long interview with host Nathan Labenz, Sentinel Bio's Joshua Monrad says AI-enabled biological attacks are held back mainly by lab robotics, not intelligence, and that knowledge barriers are eroding. Monrad backs DNA-synthesis screening and customer checks, and says Monrad would press a pause button on frontier AI for biology.

Why it matters

The same thing, for different readers.

What it means for you

Joshua Monrad of Sentinel Bio says an AI-driven bio attack is unlikely soon, mainly because robots cannot yet run a wet lab, but expects that to change in the 2030s.

For lawmakers

Monrad says laws on DNA synthesis screening and customer checks should be drafted now, so they can be scaled quickly once political will arrives.

For researchers

Monrad says wet-lab uplift studies, such as a 150-person trial, are slow and costly and lag model progress, and calls for ten Active Site-style groups.

The question it raises

Monrad says the same biology that cures disease is dual use, so banning risky tools would also cripple defence, leaving governance of data, materials and services.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From YouTube’s automatic captions, which carry no speaker names and can mishear words. Where the speaker is known only from context, the quotation says so.

“I think it's clear the models are misaligned or not well enough aligned that we could have something like an incident.”
Joshua Monrad, Co-founder and managing director, Sentinel BioMonrad, answering Nathan Labenz, on why a no-humans-in-the-loop bio incident is limited by robotics rather than by the models' intelligence or alignment.
Found word for word in the sourceHear it at 11:38 ↗
“I just genuinely think that uplift is sort of a lagging indicator. Again, these studies take a long time to conduct and then they take a long time to release”
Joshua Monrad, Co-founder and managing director, Sentinel BioAsked why uplift studies show little effect while models troubleshoot lab work well, Monrad says the studies trail model progress.
Found word for word in the sourceHear it at 37:18 ↗Speaker known from context, not confirmed by listeningspoken clean-up: i (said twice)
“I think it's better than actively training on it in the first place, but uh if the data is just like readily accessible, you haven't really caused that much friction for attackers.”
Joshua Monrad, Co-founder and managing director, Sentinel BioOn leaving sensitive data out of model training, which Monrad says helps little if the data is easy to fine-tune back in.
Found word for word in the sourceHear it at 43:48 ↗Speaker known from context, not confirmed by listeningspoken clean-up: uh
“I would absolutely if I were holding that button, I would absolutely press it.”
Joshua Monrad, Co-founder and managing director, Sentinel BioAsked by Labenz whether as benevolent dictator Monrad would pause frontier AI development until bio defenses are in place; Monrad speaks about biology.
Found word for word in the sourceHear it at 76:56 ↗
“how can we make it so that the vast majority of providers of, you know, north of 80% of providers of certain materials and services such as synthetic DNA or cloud labs are both doing the kind of KYC that would allow them to refuse an illegitimate uh query”
Joshua Monrad, Co-founder and managing director, Sentinel BioMonrad describes Sentinel Bio's win condition for materials and services providers.
Found word for word in the sourceHear it at 79:03 ↗Speaker known from context, not confirmed by listeningspoken clean-up: uh

The record

Questions
How is AI used in attacks, defence and war? · Are AI companies putting stronger safeguards into practice?
People
Joshua Monrad (Speaks, Co-founder and managing director, Sentinel Bio) · Nathan Labenz (Speaks, host, The Cognitive Revolution) · Helen Toner (Mentioned)
Also covered by
The Cognitive Revolution
What we read
We read the whole source.
Found by
The Radar, run of November 6, 2026, via yt-the-cognitive-revolution-youtube (tier 1: the source’s own channel or an official record)
Cite this
The Cognitive Revolution, “KYC Becomes KYA: Sentinel Bio's Joshua Monrad on DNA Screening, Uplift Studies & Agent Monitoring”, November 5, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/cognitive-revolution-2026-11-05-monrad-bio-risk-kya
Our notesYouTube auto-captions, no speaker names; Monrad is the guest and answers Labenz's questions, but speaker labels are inferred from context. The captions include sponsor reads (one for Anthropic) and a closing song, which are not part of the interview. The captions render the incident as 'Open Face' and mis-transcribe some names (e.g. 'UK AC', 'Kimmy'); entry uses the OpenAI Hugging Face incident story on the reading that this is what the speakers mean.