Share Your Ideas for AI Legislation
Submissions due March 29
AI policy has been a hot topic in Washington. Since the arrival of ChatGPT in late 2022, we’ve seen a proliferation of hearings, AI Insight Forums, executive action, and agency policy around AI. AI policy is one of the most rapidly evolving areas of government action, moving quickly across most of the executive and legislative branches.
The obvious exception is in legislation: none of the many proposed frameworks and bills on AI policy have passed in Congress. This can be explained in part by our divided legislature, with narrow and opposite partisan majorities in the House and Senate. However, the challenge of AI governance is not going away, and will in all likelihood increase in magnitude in the coming years. Congress cannot afford to cede its legislative power on AI.
The time is ripe for AI-related legislation. AI legislative policy has been largely bipartisan to date, with most notable legislation introduced with both Republican and Democratic sponsors. And lawmakers across the political spectrum recognize that AI is an urgent and important challenge that Congress cannot leave unaddressed.
The Federation of American Scientists is hosting an AI Legislation Policy Sprint to identify, develop, and publish a set of AI-focused policy ideas that could be implemented by Congress. The primary aim of this project is to identify recommendations for legislation that: 1) are bipartisan and pragmatic, and 2) robustly address AI-related risks and potential benefits.
What is the AI Legislation Policy Sprint?
A policy sprint is a process that guides participants as they develop an initial idea into a tailored, actionable set of policy recommendations. The AI Legislation Policy Sprint will center on sourcing legislative ideas for AI in terms of governance and regulatory models, innovation, and trust and safety. Please see the FAQ below the application form for more details.
Selected participants will have a chance to develop their ideas with guidance from policy advisors, meet with veteran policymakers to learn more about the nuances of policy implementation, and hone their ideas into an actionable AI legislative package. We are also offering a $1500 stipend per accepted proposal, in recognition of the value of our participants’ time.
What’s the outcome?
Action-ready proposals that are usually about 1,000 – 3,000 words in length. We encourage you to familiarize yourself with our library of proposals to get a better sense of the end-product. Note that there is a difference between the outputs of this sprint and past memos—they should be legislative proposals rather than memos targeted at federal agencies.
Who should apply?
For the AI Legislation Policy Sprint, we are looking for subject matter experts, ideally with policy experience or at least interest in policy engagement. We seek those who have an idea to legislatively address AI-related risks and benefits and are highly motivated to leverage the tools, skills, and networks we provide to drive this change. Ideal candidates will have deep expertise on or exposure to AI-related policy issues. For example:
- Academic researchers, including graduate students and postdocs, who specialize in AI/ML, law, ethics, sociology, or public policy and have published work related to AI governance.
- Policy analysts with expertise in AI and emerging tech policy.
- Practicing lawyers who focus on AI, privacy, intellectual property, or regulatory law.
- Technologists and data scientists who have worked on ethical AI, algorithm auditing, or related topics.
- K-12 teachers or administrators, college professors, and nonprofit executives focused on equitable deployment of AI technologies in schools.
- Activists and civil society leaders concerned about bias, transparency, and accountability in AI systems.
- Labor organizers focused on the impacts of technology and automation on the workforce.
- Entrepreneurs and business leaders developing or adopting AI in their companies.
- Journalists who cover AI technology and industry trends, especially policy implications.
This list is non-exhaustive; if you have an idea you want to share with us, please apply, no matter your background. Individuals with identities underrepresented are especially encouraged to apply. The sprint is designed around working professionals, with a time commitment of about 3-4 hours/week.
What kinds of AI legislation ideas are we looking for?
We are seeking a diverse range of policy proposals, ready to inform the Congress’s AI agenda. Some key focus areas and questions we are interested in:
- AI in Education
- What can Congress do to support the use and development of AI for educational purposes?
- What actions should Congress take to ensure that AI models used in schools do not infringe on the rights and safety of students and educators?
- What funding should Congress invest in education research and development (R&D) focused on AI that aims to improve student experience and outcomes?
- Governance & Regulatory Models
- What should we expect of AI systems employed by government bodies? To what standard should agencies hold their procurement and deployment of AI technologies?
- How can we improve the capacity of the government to encourage beneficial use of AI while reducing its risks?
- To what extent can or should we codify existing AI policy into federal legislation?
- How can we support the federal offices doing crucial AI-related standard-setting, export enforcement, consumer protection, civil rights protection, grant-making, etc.?
- Innovation
- How can we advance beneficial use of AI in fields such as education, clean energy, and healthcare?
- How should we address funding for “trustworthy” AI research?
- How can we address potential workforce and educational impacts of widespread AI adoption?
- Trust and Safety
- How can Congress help strengthen data privacy protections as AI becomes increasingly integrated into daily life?
- How can we ensure student privacy is protected in schools, while also allowing for crucial research that builds our understanding of best educational practices?
- How should Congress approach issues of:
- Algorithmic discrimination
- Democracy & disinformation
- Cybersecurity
- The use and development of dual-use foundation models
- How can Congress help strengthen data privacy protections as AI becomes increasingly integrated into daily life?
We will review applications after the close of the application window on March 29. We anticipate informing applicants of our decision by April 10 and kicking off the sprint in mid-April.
Interested parties begin by completing our official submission form. Each submission requires prospective authors to submit information on who they are, as well as a summary of their idea to help our team understand and evaluate each submission. We recommend reviewing the form in advance and thinking through each answer to the questions before officially submitting your proposal. We also suggest reviewing existing papers on our website for examples of previous work.
If selected, participants will be expected to attend up to 3-4 sessions of programming during the sprint, with additional asynchronous time spent drafting their legislative proposal. Sprint participants receive comprehensive support from the FAS team that includes:
- One-on-one mentorship from FAS staff members who will serve as a resource during the proposal development and refinement stages.
- Access to technical editing and other resources to ensure a compelling final written product.
- A $1500 stipend per accepted proposal, in recognition of the value of our participants’ time.
Following the completion of the sprint, we aim to invite authors to pitch their proposal virtually to policymakers and congressional staff. FAS will work with our legislative partners to advance participants’ proposals; however, we cannot guarantee that any specific recommendations will be taken up.