New Ethereum talks, every Monday. The week's conference uploads by event, in your inbox.

Loading player…

Biorisk as Common Resource: The Critical Digital-to-Physical Interface – San Francisco 2025

Funding the CommonsThu, Oct 9, 2025, 12:00 AM

Biorisk as Common Resource: The Critical Digital-to-Physical Interface: What kind of public good is biological risk—and how do we decide what to prioritize? Max Langenkamp offers a clarifying lens: the digital-to-physical frontier. From standardizing KYC for DNA orders to improving synthesis screening tools, he outlines concrete steps for reducing the risk of synthetic biology misuse. As design tools become more powerful and accessible, the greatest threats may come not from hypothetical super-pathogens, but from overlooked infrastructure gaps. Speaker: Max Langenkamp (SecureDNA) This session was part of the 12th edition of Funding the Commons, held March 15–16, 2025, at the Internet Archive in San Francisco. Learn more at https://www.fundingthecommons.io

Transcript

What kind of a common resource is biological risk and how do we prioritize what to pay attention to in this broad landscape? This is what I'm going to be talking about for the next five minutes. Uh I'm going to be arguing for a heristic that I think is really clarifying in the space of biological risk which is the digital to physical frontier. But before I continue just a brief introduction. My name is Max Langenampamp and I uh lead the hardware and policy research at security and here's a picture of one of my mentors Ron Revest who some of you may recognize for his role in public key cryptography.

Uh also a brief note on SecureDNA is SecureDNA is a nonprofit that provides a free and privacy preserving verifiably up-to-date DNA screening system. So if you're someone like uh Twist uh someone who provides DNA to people uh we have a tool where you can see if your customers are trying to order smallox or influenza. But back to a more general question uh about what we're doing here. I like this graph which is from Michael Nielsen essay to generally consider the role that technologies play in enhancing uh abilities and and enabling harms. The rough thing that I see uh the the mandate for someone like me is uh to move this this uh the trajectory of of possible harm up to make it so that technology is is helpful without uh threatening our civilization as a whole.

Now when it comes to biocurity there's a lot of uh there are a lot of things that you can be concerned about in addition to the whole world of of spillover uh of of natural cause pandemics if you do some of the threat modeling the most worrying things from my perspective are the deliberate misuse cases. So this can come from actors that are at the nation state level, but also at the groups group level and and even in the individual and potentially autonomous agent level. Here's some examples of scenarios that people have have worried about. Maybe your new LLMs can help a disgruntled grad student make a bioweapon. Maybe there's a new biological design tool where you can make a custom pathogen.

And here's some here are some other things too. uh some other examples that people have uh worried about. But the question that I want to to pose is even as there's so many of these different scenarios, how do we prioritize and make sense of what to do in this space? How do we defend against deliberate misuse? For me, one of the more helpful clarifying things is to remember that at some stage when you're trying to synthesize something new, there has to be a moment when it goes from bits to atoms, there's a digital to physical interface.

And so among all of these uh among all of these more theoretical uh thought experiments of you know, can your AI make like a designer pathogen? It's often the more mundane things that have lowerhanging fruit. So secure DNA for instance focuses on on DNA screening um which is on the more technical end. Given a string of AC's, G's and T's, can you tell it has something dangerous? But I think some of the lowest hanging fruit is actually things like better KYC for DNA orders.

It's actually a fact that uh you know there is no standardized way to order DNA from a DNA synthesis provider. Everyone kind of does it ad hoc in their own way. Um and another you know lowhanging fruit uh is is physical security for labs. Um but so I'm I I have about a minute left. The thing that I'll plug and I'll emphasize is that the DNA screening part is I wouldn't say it's a solved problem but there's been huge progress.

So there's many different in addition to security there's people like ACL or the common mechanism from NTI that have put a ton of effort into developing good screening tools. Right now we're missing something pretty mundane which is better tools for identifying customers and knowing who has permissions and there's a lot of uh lowhanging fruit on the coordination front. It's kind of a social problem right now. So, if you're interested in hearing more about this and supporting SecureDNA, you can uh feel free to talk to me or um yeah, I should have put my email up, but max lingamp secured.org is my is my email.

But thank you. Thank you everyone. [Music] [Applause]

Automatic transcript — names and jargon may be misspelled.