Thread by @Ayjchan

New perspectives piece in ⁦@NatureComms
Wonder what all these scientists could be talking about...
⁩“actors may be incentivized to be reckless if they believe they are unlikely to be held accountable for any accidents arising from their actions.” nature.com/articles/s4146…

"A key security challenge involves attribution: determining, in the wake of a human-caused biological event, who was responsible."

Any chance this is about COVID?

nature.com/articles/s4146…

"if an incident occurs.. near laboratories working on the causative agent, there is a greater chance of it being attributed to an accidental release."

Wonder which recent incident occurred near a lab working on the causative agent. nature.com/articles/s4146…

On "technical forensics: the properties and characteristics of the agent that caused.. outbreak may provide clues as to who made it and/or who was responsible for releasing it" - recommend reading @ScolesSarah article on the limitations of this approach. futurehuman.medium.com/how-do-we-know…

tldr technical forensics can be severely limited because scientists often do not publish their discoveries in a timely manner, which is exactly the situation we are in re: covid and the type of situation we anticipate in the future.

“Everyone who’s trying to engineer knows just as much as the people who’re trying to detect engineering,” says Plant. “So you have, functionally, an arms race. You’re never going to have completely the ability to detect something that’s engineered.”

Why would scientists, especially those engaging in covert research, start publishing all their pathogen data in a timely manner if they know that this can be used to hold them accountable for accidental leaks in the future?

It's fine to keep funding technical forensics R&D, but this seems, imo, to be a sizeable misunderstanding of why scientists can't tell even today where SARS-CoV-2 came from, nature or lab-based scenarios - if that is even what this article could possibly be alluding to.

The reasons why we cannot pinpoint whether SARS-CoV-2 came from a lab is not because our Machine Learning or other tools to detect traces of engineering are underdeveloped.

The reasons are (1) it is so damn difficult to get access to information regarding even research that was funded by the US for years. Case in point: RaTG13, collected in 2013, inaccurately reported sample history and data in Jan 2020, addendum in Nov 2020. nature.com/articles/s4158…

This is not at all a technological issue. This is a transparency issue. A human issue. No amount of machine learning is going to overcome this barrier.

(2) Politics in science, conflicts of interests, an aversion among scientists to engaging in politics.

Many scientists won't even voice the possibility of lab origins publicly. The majority of the public and scientists believe the scientific consensus is covid=natural.

How can scientists across different disciplines openly and rigorously analyze natural vs lab origins when, right from the start, one entire category of hypotheses is disparaged as harmful conspiracy theories?
usrtk.org/biohazards-blo…

Imo, ensuring an open, public platform for scientists (established and trainees) to analyze origins data/evidence is much more crucial than more technologies to detect engineering - which again is not at all the key to solving the origins of covid.

The solutions to this challenge are again not better forensic tech, but human ones. There needs to be a better way of identifying conflicts of interest other than self-declaration. There needs to be open editorial+peer review, open tracking of funding (organization, individual).

On a related note, (3) the ability to assemble and enforce, without delay, an international investigation by a team free from pre-existing relationships or interests that can compromise the credibility of the investigation.

https://twitter.com/Ayjchan/status/1331665672661848069

Troubling that more than a year has passed since covid emerged and no international team of investigators has received permission to even go to ground zero or talk to doctors (reportedly silenced) or index patients. US journalists expelled from China. inews.co.uk/opinion/china-…

If another pandemic arises in the future under similar circumstances (no clear traces of engineering), my bet is that its origins are not going to be predicted by ML.

Advocate the powers that be to invest in better research tracking + accountability & investigation mechanisms.

Sometimes there's a temptation to think that if we throw enough millions of $$$ at technological solutions, we can solve the greatest challenges in the world.

But I don't think we are going to be able to machine learn, AI, or blockchain our way to the origins of covid.

Afaik internet sleuths have done more than any cutting edge technology or international investigation team to uncover key facts pertaining to the origins of Sars2. Story by @emmecola

https://twitter.com/ayjchan/status/1334870461923127297