Steph Batalis

CSET's Recommendations for an AI Action Plan | Center for Security and Emerging Technology In response to the Office of Science and Technology Policy's request for input on an AI Action Plan, CSET provides key recommendations for advancing AI research, ensuring U.S. competitiveness, and max...

The plan also promotes and emphasizes the importance of scientific, including biological, datasets- in line with @csetgeorgetown.bsky.social recommendations for the plan, which you can read here: cset.georgetown.edu/publication/..., and with other CSET work: cset.georgetown.edu/publication/....

How to stop bioterrorists from buying dangerous DNA The companies that sell synthesized DNA to scientists need to screen their customers, lest dangerous sequences for pathogens or toxins fall into the wrong hands.

Focusing on bio, one provision is a federal funding requirement for DNA synthesis screening- a useful tool in the toolbox for limiting biological risk.

Check out @stephbatalis.bsky.social and I's piece breaking down the kind of decisions screeners have to make: thebulletin.org/2025/04/how-...

Canceled grants get the spotlight at a Capitol Hill 'science fair' On Tuesday, scientists held an event organized by House Democrats in which they stood in front of posters outlining their work — and the federal cuts that now threaten it.

On Tuesday, scientists held an event organized by House Democrats in which they stood in front of posters outlining their work — and the federal cuts that now threaten it.

Antimicrobial resistance is a huge issue and an oft-forgotten killer. It kills more people each year than HIV/AIDS or malaria.

This article is fascinating- it points out that while much of the AMR prevention discussion focuses on overuse of antimicrobials, underuse can also be a major issue.

Amidst all the discussion about AI safety, how exactly do we figure out whether a model is safe?

There's no perfect method, but safety evaluations are the best tool we have.

That said, different evals answer different questions about a model!

"Red-teaming" isn't a catch-all term (or methodology!) to evaluate AI safety. So, what else do we have in the toolbox?

In our recent blog post, we explore the different questions we can ask about safety, how we can start to measure them, and what it means for AIxBio. Check it out! ⬇️

America’s response to measles is eroding its ability to deter biological attacks The rising death toll for a preventable disease reveals just how ill-prepared the country is to handle a malicious bioweapon.

As I wrote in Defense One, "Dismantling critical preparedness offices, cutting infrastructure and funding, and allowing misinformation to derail the response are not just bad for healthcare—they’re dangerous national security signals." www.defenseone.com/ideas/2025/0...

News like this isn't just a concern for public health practitioners - it should also be a big red flag for U.S. national security folks.

America's biodefense strategy uses robust health infrastructure to deter bad actors. Right now, we're tearing down our own defenses so adversaries don't have to.

Trump's Gain-of-Function Order Prompts Questions It comes as another guidance was set to take effect

A Trump executive order calls for putting an end to "dangerous gain-of-function research" -- prompting experts to share concerns about its potential impact on infectious disease research. @stephbatalis.bsky.social @raz524.bsky.social
www.medpagetoday.com/special-repo...

This has only become more relevant in the past week. Research cuts, new hurdles for vaccine trials, and tariffs on key medical countermeasures aren’t just risking Americans' health—they’re dismantling U.S. biodefense so our adversaries don’t have to.

@defenseone.bsky.social

The Pandora Report is easily one of my favorite newsletters right now. A great one-stop shop for timely health + biosecurity policy news, especially with too many “Breaking News” alerts flying around fight now to follow. Plus, it always sends me down rabbit holes with great links. Highly recommend.

To be very clear: the U.S. should prioritize public health for its own sake. Every person deserves access to systems that protect them from preventable disease.

But it’s also true that national security depends on a strong public health foundation to detect, contain, and treat dangerous outbreaks.

America’s response to measles is eroding its ability to deter biological attacks The rising death toll for a preventable disease reveals just how ill-prepared the country is to handle a malicious bioweapon.

🚨 Latest op-ed is out in Defense One!

“Dismantling critical preparedness offices, cutting infrastructure and funding, and allowing misinformation to derail the response are not just bad for healthcare—they’re dangerous national security signals.”

www.defenseone.com/ideas/2025/0...

Thanks for sharing, Emmy!

How to stop bioterrorists from buying dangerous DNA The companies that sell synthesized DNA to scientists need to screen their customers, lest dangerous sequences for pathogens or toxins fall into the wrong hands.

When customers order synthesized DNA, how do companies know whether it's safe to send?

For @thebulletin.org, CSET's @stephbatalis.bsky.social & @vikramvenkatram.bsky.social explore what it takes to keep things safe. thebulletin.org/2025/04/how-...

In our latest piece for @thebulletin.org, @vikramvenkatram.bsky.social and I put YOU, the reader, in the shoes of a 🧬DNA synthesis provider🧬 to demonstrate how much tougher customer screening is than it may seem, and why guidance would be helpful for providers trying to make tricky decisions.⬇️

Public Statement on Supporting Science for the Benefit of All Citizens TO THE AMERICAN PEOPLE We all rely on science. Science gave us the smartphones in our pockets, the navigation systems in our cars, and life-saving medical care. We count on engineers when we drive acr...

"We all rely on science [...] Businesses and farmers rely on science and engineering for product innovation, technological advances, and weather forecasting. Science helps humanity protect the planet and keeps pollutants and toxins out of our air, water, and food."

docs.google.com/document/d/1...

Opinion | Funding for R&D isn’t a gift to academia Investing in scientific research and development is vital to U.S. security.

NEW: Sen. Todd Young (R-Ind.)& Trump's former NatSec Advisor Matt Pottinger make the case that funding for scientific R&D isn't a gift to academia. It's vital to U.S. national security, as China, the United States' primary strategic adversary, is already investing heavily in R&D.
shorturl.at/PgUK0

Not only this, but most of the NIH research was to ID biological targets/causes of disease rather than direct drug development—work that often isn't in the domain of the private sector. Cutting federal funding would mean less of the foundational science that makes future breakthroughs possible.

A white man in a blue jacket holds a sign that reads “I’d rather be in lab”

A beautiful day and the scientists just wanna science, man.

Federal funding underpins American research across 'hot' AI + bio research clusters – Emerging Technology Observatory As China challenges, NIH and NSF funds play a key role in U.S. research efforts

TLDR: We can’t try to gain global leadership *and* destabilize our NIH/NSF funding edge.

We aren’t just at risk of losing out on essential scientific advances. The global leader also gets to set global norms, shape future tech, and a HUGE economic advantage.
eto.tech/blog/federal...

I was asked on a panel yesterday what keeps me up at night: It’s this.👇 Even though my day job is to think about bioweapons + bad actors, THIS is what has me fearful. The long tail of this could undermine bio/med innovation, public health infrastructure, and the economies built on them for decades.

At a time of intense global competition, reducing funding to NIH and NSF isn’t just a budget issue. It’s a strategic misstep that could undermine U.S. influence in biotech and AI for decades.

Read more in my recent post with CSET's ETO👇

Overall, Evo 2 has captured well-deserved attention. It has me excited to see what's next in the AIxBio space...and curious about how AIxBio policy will continue to evolve with new cutting-edge advances.

3. NVIDIA’s announcement simply calls Evo 2 a "foundation model," highlighting bio/policy/developer differences on shared terminology. To a biologist, Evo 2's range *is* broad-purpose. If policies aren't meant to capture bio AI, that needs to be explicitly stated in the regulatory language.

2. Even without the EO, Evo 2’s team addressed biosecurity - mainly by removing viruses that infect eukaryotes from training data. However, those sequences could be re-added since the model is open-source, and it's unclear whether they red-teamed other biological threats like bacteria or toxins.

1. With the Biden EO on AI now revoked, Evo 2's massive size would have triggered the EO’s reporting requirements. Trained with an estimated 2.25 x 10^24 FLOPs, it far exceeds the EO’s 10^23 threshold for biological models. But now, no such obligation exists.

Evo 2's release yesterday isn’t just impressive (though it is!)—it’s also making waves in the AIxBio policy space as the largest fully-open biological AI model to date. 🧬

⬇️Here are a few policy implications I’ll be keeping an eye on:

My favorite thing about working at @csetgeorgetown.bsky.social? Learning something new every day from the experts I'm lucky enough to sit next to - from cybersecurity to AI evals to export controls (and more!)

Don’t take it from me - join the virtual CSET water cooler here: bsky.app/starter-pack...

The dogs of Dupont Circle in the snow this morning! @capitalweather.bsky.social #dcsnow #dmvsnow #dcsnow2025 a