We’re interested in funding projects that can be deployed at scale to significantly reduce exposure to pathogens (especially airborne ones). Our goal is to build resilience to both respiratory pandemics and scenarios where pathogens are pervasive in the environment (e.g. mirror bacteria) through reducing pathogen inhalation by orders of magnitude. Promising projects could involve developing new tools or increasing the likelihood that existing transmission suppression technologies can be used on short notice (e.g. by strengthening the evidence base or by identifying scalable deployment strategies). We are particularly interested in cost-effective and highly scalable solutions for respiratory protection, air disinfection, and tools for verifying the efficacy of those solutions.
We’ve previously supported:
- ProEquip, an organization working on stockpiling respiratory PPE for critical infrastructure workers
- Blueprint Biosecurity, a nonprofit that develops and advances transmission suppression technologies, such as PPE, far-UVC, and glycol vapors
Hypothetical, non-exhaustive list of example projects we might be interested in funding:
- Testing methods to replace filters with common household textiles (e.g. making an effective Corsi-Rosenthal box without HVAC filters)
- Developing a very cheap, open-source device to accurately monitor particle concentrations inside respirators
- Gathering evidence on the ability of drug-free nasal sprays (e.g. hypertonic saline or carrageenan gum) to reduce viral load in the upper respiratory tract
- Developing DIY-capable protocols to measure microbe concentration reductions from UV, glycols, or air filters in a space
- Designing and testing easy-to-stockpile or easy-to-manufacture room air disinfectant units (e.g. an air freshener gel that can passively disinfect occupied spaces) in a variety of building types
- Engaging philanthropists to support country-level PPE stockpiles for key workers
We’re interested in funding projects that reduce the risk of terrorists, states, or AI agents acquiring powerful biological weapons.
In some cases, this involves governance of specific high-risk technologies like mirror life. More commonly, it involves efforts to reduce the risk of AI systems being used to help malicious groups develop biological weapons, whether by acting directly on AI systems themselves and or on relevant technologies (like nucleic acid synthesis). As AI capabilities advance rapidly, including in areas relevant to dual-use biology, there is an urgent need to evaluate dangerous capabilities, build safeguards, and develop institutional infrastructure before a serious incident occurs.
Among grants addressing AIxBio specifically, our primary focus is on projects that (1) generate rigorous AI evaluations credible enough to unambiguously prompt action if they trigger, and (2) develop and red-team safeguards that can help detect or deter misuse and remain robust even as AI coding capabilities advance.
We’ve previously supported:
Hypothetical example projects we might be interested in funding:
- ‘Rule-in’ evaluations testing whether AI models can assist experts at novel dangerous pathogens — similar to recent studies done on LLM-assistance for novice performance
- An organization to regularly red-team whether a well-resourced group could circumvent current misuse safeguards by frontier AI companies, such as jailbreaks, know-your-customer, or access controls
- Work to improve LLM classifiers’ ability to detect misuse, and other proposals to better distinguish between legitimate and harmful users, such as tiered access
- Using AI to accelerate biosecurity defenses, such as patching vulnerabilities in DNA synthesis screening, know-your-customer policies, and other safeguards
- Work to investigate whether safeguards could be put in place such that AI bioweapon capabilities become harder to distill or jailbreak than other safeguards
- Empirical analysis of how quickly open-weight models are catching up on specific bioweapon capabilities, how large a role distillation plays, and how likely trends are to continue
- Empirical analysis of whether specific biological datasets are genuine bottlenecks for dual-use capabilities in biological AI models or could be overcome via more powerful models and compute
We’re interested in funding work that helps to educate policymakers and the public about risks and potential interventions in many of the areas listed above, and to help governments better prevent and prepare for biological catastrophe.
In this area, we have previously supported the Mirror Biology Dialogues Fund on their work with top scientists around the world to coordinate a response to the risks of mirror life, Blueprint Biosecurity’s 501(c)(4) arm and policy advocacy work, and more.
Hypothetical example projects we might be interested in funding:
- Work to increase policymakers’ awareness of catastrophic risks from the deliberate misuse of biology
- Advocacy for government stockpiling of elastomeric respirators and other reusable PPE to protect vital workers in the event of a catastrophic biological event
- Advocacy for comprehensive, enforceable gene synthesis screening regulations and know-your-customer laws in the US, EU, and other key jurisdictions
- Development of legal and institutional frameworks for biosecurity threat reporting (analogous to models like NCMEC for combatting child exploitation)
- Advocacy for sustained government funding of pathogen-agnostic detection infrastructure
- Policy frameworks for governing mirror biology and dual-use AI in biology, and analysis of regulatory pathways for rapid deployment of transmission-suppression technologies in crises
- International dialogues to foster governance to ensure mirror life remains out of reach
- Approaches to communicate to broader audiences about the risks of mirror life (including policymakers, the public, scientists)
- Advocacy for widespread deployment of portable air cleaners
- Work on how to detect and disrupt bioterrorists
We’re interested in funding activities that help talented people from various backgrounds pursue work focused on addressing global catastrophic biological risks, support the work of those already in the field (e.g. via events that help build professional networks), or advance the public discourse around risks and opportunities in this space.
Outside of our Career Development and Transition Funding program, some examples of field-building initiatives we’ve funded in the past:
We think field-building activities are essential drivers of progress across the space. Hypothetical examples of programs we might be interested in funding:
- Incubator/co-founder matching program for high-priority biosecurity projects
- Fellowship program at a top university that serves as a biosecurity “on-ramp” for computational biologists
- Training program focused on discussing the risks of mirror life and understanding/developing policy solutions
- Investigative article series at a reputable journalistic outlet
- Biosecurity side-event/happy hour connected to a major conference in a related area, like microbiology or indoor air quality