Bio x AI Policy Development Sprint
The Bio x AI Intersection Needs Stakeholder Engagement To Promote Innovative Governance
Congress is actively working on artificial intelligence (AI) governance, mainly focused on social implications of AI use, however approaches to governance for AI’s intersections with the biosciences must also be tailored to the needs of the scientific community. AI, machine learning, and deep learning have been utilized in the biosciences since the turn of the century and it will be crucial to create governance that allows for the mitigation of biosecurity risks associated with the technology while advancing R&D innovation.
The current attention on AI and the upcoming interim report to Congress by the NSCEB provide an important policy window and acts as a call to action that requires stakeholder input in order to create governance and policy recommendations that enable innovation while mitigating risks. The Federation of American Scientists are crowdsourcing policy input to create timely, well considered policy recommendations that address the key questions listed below to enable safe implementation and governance of AI in the biosciences.
What We’re Looking For
AI or bio experts within academia, the biotech industry, an AI lab, or other non-governmental organization that have policy ideas, input, or recommendations around the following focus areas:
- How can we implement oversight on current and future bio-related tools that utilize AI enabled technologies (e.g., AlphaFold2, etc) in order to mitigate biosecurity risks associated with the technology while advancing R&D innovation?
- Are there other ways to reduce the potential for misuse with these technologies?
- AI model training requires an immense amount of data and various data sets (e.g., protein structure databases, genomic sequence databases, high-throughput data bases, etc). How should we address the need for scientists to have access to and ability to generate these types of data sets in order to train bio-related AI tools while also balancing the potential for misuse of that data?
- How can bio-related AI or ML tools be applied to or reworked in order to reduce the risk of misuse or to improve biosecurity more broadly?
- Identifying the impact that AI enabled self-driving labs would have on the biosciences and what biosecurity considerations should be made around them
Other Areas:
- What is the difference between ML and AI in the biosciences, would concise definitions for each or an agreed upon framework guide the use of AI technologies in the biosciences?
- How does scale and access to information affect these questions?
- Identifying how AI-enabled bio should be handled within the broader context of oversight of biosciences and biotechnology. (E.g., DNA synthesis screening; sharing of pathogen datasets).
- Identifying how biotechnology-related AI should be handled within the broader context of oversight of AI
- Standards for Safe Open Source Technology
How it Works
- You submit an idea below that matches one or more of the relevant prompts (listed under “What We’re Looking For”). Submissions should be no more than 350 words and should include a clear plan of action in their recommendations.
- A selection of the most compelling ideas will get to work with our team of experts to craft policy recommendations to be delivered to the NSCEB.
- Authors with promising ideas will be invited to participate in the sprint and work with other subject matter experts to refine their policy ideas to actional policy recommendations.
Ultimately, FAS will synthesize recommendations with contributors and collaboratively provide inputs to the policy leaders working with and within the White House, Congress, and federal agencies on promoting AI driven innovation in the biosciences while mitigating associated risks.
Have more questions? See the FAQ below or get in touch with our team at fas@fas.org.
**Note that FAS will be mindful of information hazards and will not publish any information about specific tools or vulnerabilities that could inform anyone aiming to misuse these tools**
We ask that all submissions follow the format above and feature a clear and concise summary and plan of action so that our teams may quickly understand the main thrust of your recommendations. You are welcome to link to relevant reports or data within your submission.
We are looking for ideas that are a) specific and grounded in the current Bio AND AI landscape, b) are actionable at the federal level, and c) represent opportunities for bipartisan action. Creativity is also encouraged!
There will be ~3-5 group meetings to discuss the Bio x AI landscape, to identify the challenges and opportunities in the field, and to refine the ideas that you have already submitted, as well as craft new policy recommendations based on group conversations. After that, you will write a 1-1.5 page document outlining the challenge, opportunity, policy recommendation, and plan of action which we will put together to create a menu of recommendations. We estimate this will take anywhere from 15-20 hours over two months.