Panoplia Laboratories (now operating as Active Site) is a nonprofit that evaluates the risks and capabilities of AI-driven biology through wet lab research, and develops broad-spectrum antivirals for pandemic preparedness.
Panoplia Laboratories (now operating as Active Site) is a nonprofit that evaluates the risks and capabilities of AI-driven biology through wet lab research, and develops broad-spectrum antivirals for pandemic preparedness.
People– no linked people
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- $854,335
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26Panoplia Laboratories, Inc. (now operating as Active Site) is a 501(c)(3) nonprofit organization founded in June 2023 by Brian Wang and Alex Kleinman in Cambridge, Massachusetts. The organization was originally established to develop broad-spectrum antivirals using synthetic biology tools, aiming to create medicines deployable at the start of a pandemic before traditional vaccines could be developed. The founding team came from related biosecurity work. Brian Wang, who holds a PhD in organic chemistry from UC Berkeley, previously conducted postdoctoral research in synthetic biology under Kevin Esvelt at the MIT Media Lab and served as Head of R&D at Alvea, where his team designed, produced, and tested RNA vaccine candidates against COVID-19 and influenza. Alex Kleinman also previously worked on the science team at Alvea. In its early phase, Panoplia pursued the development of inhalable, DNA-encoded broad-spectrum antiviral proteins called virus-targeting chimeras (VIRTACs), inspired by the innate immune system. The organization built capacity for in-house animal studies and conducted pilot studies testing antiviral candidates against influenza. In 2025, the organization underwent a significant strategic pivot and rebrand. Brian Wang departed to become a Programme Director at the UK's Advanced Research and Invention Agency (ARIA), and Joe Torres, the organization's lead scientist with a PhD in Molecular and Cell Biology and over 10 years of experience in immunology and virology, assumed the role of Executive Director. Under this transition, the organization rebranded to Active Site and shifted its primary focus to empirically evaluating the risks and capabilities of AI-driven biology through wet lab research. Active Site's flagship project was a landmark randomized controlled trial conducted over the summer of 2025, funded by a $600,000 grant from Sentinel Bio and the Frontier Model Forum. The study recruited 153 novice participants and randomly assigned them to either an LLM-assisted group or an internet-only control group, then measured their performance on fundamental wet lab molecular biology tasks over 8 weeks, including procedures simulating viral reverse genetics work. The results, published as a preprint on arXiv, found that while AI showed signs of helpfulness at individual steps, it did not produce a statistically significant effect on end-to-end task success, surprising many domain experts. Results were pre-reviewed by an independent advisory board of biosafety and national security experts before publication. The organization has also collaborated with the Forecasting Research Institute (FRI) to study how well superforecasters and domain experts predicted the study's outcomes. As of March 2026, Active Site is actively hiring for multiple positions including Head of Operations, Lab Operations Manager, Scientist, and Research Associate, signaling plans to expand its research capacity.
Theory of Change
Updated 05/18/26Panoplia Laboratories (Active Site) operates on the theory that empirical, controlled experiments measuring how AI models actually augment human capabilities in biological laboratory settings are essential for understanding and mitigating AI-driven biosecurity risks. By conducting rigorous randomized controlled trials with real participants in real wet labs, rather than relying on theoretical assessments or computational benchmarks, the organization generates ground-truth data about whether and how frontier AI models enable biological capability uplift. This evidence directly informs risk assessments by AI developers, policymakers, and the biosecurity community, enabling more targeted model-level safeguards and governance frameworks. Previously, their antiviral work aimed to reduce existential biological risk by developing pathogen-agnostic medicines that could be stockpiled and deployed immediately at the onset of any pandemic, eliminating the dangerous gap before traditional vaccines become available.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.