Skip to main content
MANIFOLD
Will someone with little to no bioengineering knowledge be able to use an AI to create a bioengineered virus by 2024?
16
Ṁ330Ṁ1.6k
resolved Jun 22
Resolved
YES

https://youtu.be/lh72yLK4GQA
The main restriction on making nuclear weapons at the moment is availability of uranium. The materials for making viruses are much less rare. Will someone be able to make a bioengineered virus using an AI without having anything more than a bachelors at most in Biology?

Market context
Get
Ṁ1,000
to start trading!

🏅 Top traders

#TraderTotal profit
1Ṁ236
2Ṁ97
3Ṁ70
4Ṁ43
5Ṁ28
Sort by:
predictedYES

I'm sad to say that I'm going to be resolving this to yes. I won't share the full details as to why, but it's something similar to https://twitter.com/andrewwhite01/status/1670794000398184451/photo/1
I'm sharing the above, since it already has 450k views on twitter.

This seems incomplete - what kind of lab equipment are you assuming access to? If the AI tells you how to synthesize a virus given complete access to Merck's labs is that sufficient? Because that's a lot weaker than, say, it tells you how to synthesize a virus using parts you can buy from walmart.

predictedYES

@vluzko Things that anyone with $10,000 and an internet connection can get.

predictedYES

Without going into details, for obvious reasons, I am currently leaning towards resolving this at yes. I'm waiting to hear back from a molecular biology researcher and going to do some tests myself.

If you do find a method that satisfies your criteria, published publicly or otherwise, I think it would be best if you didn't publish any link to it here, rather just resolve this question.

predictedYES

@WXTJ yeah, of course. Same reason I wouldn't share ways to discretely make petrol bombs.

How do you resolve a market in "is someone able to do x" without actually doing it?

It seems the only feasible resolution criteria are news reports oft someone actually trying to engineer a pathogen.

@Schwabilismus also why engineering a pathogen? It should be "easy" enough to produce rhizin vor anthrax who already count as potent bioweapons.

@Schwabilismus *rhizin OR Anthrax.

predictedYES

@Schwabilismus As in, such a method posted online and verified by reputable reseachers/organizations.
Also, I know some Biomolecular researchers and if they go through the steps and say it checks out, I'd be inclined to resolve it to yes. Though, I'd want to check it with some specialists. If they say no, similarly, but towards no.

A virus that can spread rapidly, such as covid is the main focus here.

To be clear, if someone with just a bachelors in Biology is able to do all the steps needed to make a virus and all they would need to do is make it in a lab and the AI (or another AI) helps them know exactly how to do so and how to get access to a lab/how to homebrew such a virus, I would count that.

predictedYES

@KabirKumar you can actually make a bioweapon by just letting a slice of pork sit and rot for a while.