SecureBio is a research non-profit that works to advance biotechnology safely and prevent catastrophic pandemics
Joined January 2025
- Tweets 176
- Following 1
- Followers 676
- Likes 8
71 Photos and videos
Jun 10
We’ve seen LLMs beat expert virologists in multiple-choice tests.
The question now is: can they use that knowledge in practice?
Our new assessment, ABC-Bench, tests LLMs on real lab work and biological problem-solving.
1
1
7
1,084
Jun 10
Creating a harmful virus, or something similar, involves many steps, and we only assessed 3 of them.
It’s still very unlikely a random person could make a virus using ChatGPT, but our safety systems must evolve alongside AI capabilities.
1
1
226
Jun 10
In upcoming work, our researchers will continue to benchmark agentic capabilities of frontier models.
Read more on our substack: securebio.substack.com/p/how…
1
142
Jun 2
Why run AI biorisk evals at all? In his new post, @JasperGeh explains the biorisk evidence hierarchy (first-principles arguments, evals, uplift RCTs) and why evals provide the best evidence-per-dollar.
Read the full post here: securebio.substack.com/p/the…
2
6
408
May 1
Our attention to biorisks posed by AI needs to match the current attention given to cyber-risks. The staged release of Claude Mythos in order to bolster defenses in key industries is necessary to shore up resilience against a new class of cyber-risk across critical industries. We should do the same with biorisks.
Apr 29
Between Mythos, Anthropic v Pentagon, and the drumbeat of biosecurity concerns, I think we're entering a period where
(a) AI as a consumer product (chatbot coding assistant) will continue to behave like a "normal" technology—a very powerful tool that does all sorts of useful things alongside workers without imminently wiping out tranches of the labor force; while...
(b) as bio-/cyber-/national security threat, AI is becoming something else entirely, a sort of hydra of existential risk that's going puncture ppl's confidence that "AI is a normal technology" and force both the labs and the federal govt to establish rules on the fly, while open models race to catch up to the frontier
2
2
22
5,735
May 1
Bio and cyber AI capabilities increase together. Claude Mythos scored higher than any other model on CyBench, an evaluation of cybersecurity capabilities. It also scores higher than any other model (and 100% of human experts) on our Virology Capabilities Test, which measures advanced lab troubleshooting capabilities. Not only that, but Mythos showed the greatest score increase we’ve ever observed.
1
1
418
May 1
We need to take AI biorisks equally seriously while we still have time to act: pre-deployment evaluations, third-party red-teaming, and managed access are low-friction ways to better understand and manage biorisks posed by advanced AI. The trendline is clear: biological capabilities of frontier AI are increasing, and unless we improve the biosecurity of these models, the risks of bio-incidents will grow. We need to secure the downside of these models, in order to preserve their upside.
2
295
SecureBio retweeted
Apr 28
Very glad to see government officials at the highest level taking the threat of AI-engineered bioweapons seriously
Apr 25
NEWS from my @WSJopinion interview with @SecScottBessent
Treasury officials say that at the Trump-Xi summit in Beijing next month, history will be made. For the first time ever, the leaders of earth’s two greatest powers will discuss AI as an agenda item. Messrs. Trump and Xi will look for areas of mutual cooperation and explore ways to work together on security and threats from nonstate actors.
Full interview here:
wsj.com/opinion/scott-bessen…
@lizalinwsj @stuartlauchina @danstrumpf @yuanli233 @paulmozur @trippmickle @kateconger @cade_metz @evadou @christian_shep @Cat_Zakrzewski @GerryFShih @DemetriSevast @EleanorOlcott @AnnaNicolau @tabbyleung @parmyolson @daveyalba @jackiewattles @iansking
8
31
189
39,647
SecureBio retweeted
Apr 29
Glad to see AI bio-risk entering the public consciousness. It really is existential.
Apr 29
Between Mythos, Anthropic v Pentagon, and the drumbeat of biosecurity concerns, I think we're entering a period where
(a) AI as a consumer product (chatbot coding assistant) will continue to behave like a "normal" technology—a very powerful tool that does all sorts of useful things alongside workers without imminently wiping out tranches of the labor force; while...
(b) as bio-/cyber-/national security threat, AI is becoming something else entirely, a sort of hydra of existential risk that's going puncture ppl's confidence that "AI is a normal technology" and force both the labs and the federal govt to establish rules on the fly, while open models race to catch up to the frontier
7
12
115
16,994
Apr 23
We at SecureBio tested GPT-5.5’s biorisk-related capabilities: virology and pathogen knowledge, niche scientific knowledge, agentic bio capabilities, and bio AI tool usage. GPT-5.5 scores at or near the top on all of the evaluations we gave it. Some highlights:
2
12
60
6,312
Apr 23
3rd party pre-release testing like this is vital to understand how biosecurity-relevant capabilities are evolving and to ensure that safeguards, evaluations, and policy responses keep pace before these models reach the public.
1
8
797
Apr 23
Many thanks to the OpenAI team for engaging with us on pre-release testing! Our full results and methodology are described in this report: substack.com/home/post/p-195…
13
618