The proposed rules also target physical DNA manufacturing equipment.
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
Many users dismissed AI leaders urging Congress to mandate screening for synthetic nucleic acids as regulatory capture or CEO fearmongering, while some praised it as a substantive safety step.
shit is getting real
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
OpenAI, DeepMind, Anthropic CEOs back mandatory DNA synthesis screening
A coalition of AI leaders, synthesis-industry executives, biosecurity researchers, and former national-security officials published an open letter in June 2026 urging Congress to make screening and recordkeeping of synthetic nucleic acid orders mandatory in the US, arguing that rapidly improving AI is eroding the knowledge barriers that have historically kept bad actors from building biological weapons.
Signatories - including Demis Hassabis, Sam Altman, Dario Amodei, and Nobel laureate David Baker - frame screening as a well-understood, low-disruption measure already practiced voluntarily by major providers, and call for action this congressional session plus consistent state-level standards.
This is good! I started red teaming LLMs for biorisks/weapons risks in my bedroom in Nov 2022.
In 2023 a lot of people said we were overreacting -- 'models won't be better than internet search'. True then, but the important point is they were going to get a lot better.
Now I think there's clear reason for caution. The ultimate solution is inventing + deploying biodefenses. This is one step.
I fully expect humanity to conquer pandemics in the 21st century
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
WSJ coverage:
Great. Now you've told the terrorists what to buy, while not having implemented anything to prevent or trace it.
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
https://www.wsj.com/politics/policy/top-ai-ceos-call-for-law-protecting-against-biological-weapons-88f2f99f
One of my core tenants is that we need to prepare for a world where Ai capabilities are widely diffused, and this means carefully understanding other bottlenecks (than the AI models) to ensure safety.
Happy to sign this and it’s an honor to be listed by so many of my role models.
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
@logangraham >I fully expect humanity to conquer pandemics in the 21st century
We shall conquer death.
This is good! I started red teaming LLMs for biorisks/weapons risks in my bedroom in Nov 2022.
In 2023 a lot of people said we were overreacting -- 'models won't be better than internet search'. True then, but the important point is they were going to get a lot better.
Now I think there's clear reason for caution. The ultimate solution is inventing + deploying biodefenses. This is one step.
I fully expect humanity to conquer pandemics in the 21st century
Many of my mutuals are also signatories, congratulations to all of them.
revisions to existing nucleic acid screening requirements were mandated by an EO POTUS signed a year ago; I worked on them while in govt. I genuinely don’t know what happened to that work after I left but it is nine months behind schedule. Congress acting is better anyway.
@QRDL @Sean_Sooch18 From the red team report: https://red.anthropic.com/2026/mythos-preview/
@AndrewCurran_ Looks like a win for safety on paper but appears to be a result of acc in actual practice. Losing faith in agency of the safety position. Doesn't appear to be politically attainable even if it's desirable. This is ahead of schedule imo
This is obviously a good idea, but has limited shelflife.
What’s the plan for when AI can build this equipment from scratch?
Sam Altman, Dario Amodei, Demis Hassabis and many others have signed a letter urging Congress to increase security on orders of synthetic nucleic acids - and the equipment needed to make them - as models continue to become increasingly bio-capable.
@Sean_Sooch18 Yes, we are speeding up. The emergent capabilities of Mythos class+ will include bio even if they don't train for it.
Andrew, I signed, this is important work - but we need to be really careful to not get distracted from the bigger problem of rapidly decentralizing capabilities to create / deploy biothreats. Detection, identification, response - globally and persistent/pervasive - should be the top priority.
DNA synth will be distributed, it won’t be centralized in a way that can be effective with this approach. Desktop, at home, new synthesis tech coming, etc. Not to mention we can’t regulate China, etc. and this is a globally distributed technology.
This kind of legislation is sort of like telling people not to write viruses on all their home computers or, worse, if they do we will monitor them all to see what they type (which people would get around of course). Important now, not practical in medium/long run.
Necessary to reduce some risk, not nearly sufficient… nor where our top priorities should be.
Only practical thing is to make detection and response work great, global, and fast.
e.g. - The "finger in the dike" refers to the classic legend of the "Little Dutch Boy" who selflessly plugs a small leak in a dam with his finger, holding back the sea to save his town from a devastating flood. Small, temporary stopgap to a major, escalating problem.
Fuck me, this isn't a science fiction novel.
Making pathogens is not that simple and it's the weapon of choice of the dumbest, most intellectual challenged thing you could possibly use AI for.
Suicide bombers looks like geniuses next to whatever moron thinks bioweapens are a smart idea.
@AndrewCurran_ Industry collaboration on political safety-related initiatives. This is actually quite rare (maybe even unheard of).
@AndrewCurran_ @Sean_Sooch18 I'd like to see a cite saying them denying they did RL on cybersec. But even if that was the case, RLVR on coding/math likely enabled it. RLVR is not practically feasible on bio
The people building AI are asking to be regulated. That should tell you something.
AI surpassing most humans isn't an if, it's a when. And here's the uncomfortable part: every safety system we've built rests on one assumption, that a human is still in the loop. Someone orders the DNA. Someone runs the lab. That person can be checked, denied, punished. The whole letter is about guarding that one human chokepoint.
Physical AI dissolves it. The danger was never a machine that wants to harm you. It's that the friction protecting us, the fact that a person has to be willing to act, simply goes away.
We're locking the door we can still see. The next one won't have a handle.
@QRDL @Sean_Sooch18 Mythos hacking capabilities were completely emergent.
