AI Rapid Response Fellowship

AI capabilities are advancing rapidly, and the most recent generation of models has brought the security implications of those advances into sharp public focus. Policymakers and the federal government are mobilizing to respond, but have also identified a consistent bottleneck to effective action: a shortage of government staff with domain expertise in AI, cybersecurity, and other relevant fields.

The AI Rapid Response Fellowship is a focused effort to help close that gap. Horizon is recruiting a specialized cohort of exceptional technical and policy talent to work full-time at executive branch offices on AI security-related issues, with placements beginning as soon as possible.

Unlike our regular Horizon Fellowship program, which supports people transitioning into public service careers, the AI Rapid Response Fellowship is built for people with relevant prior experience who can contribute to executive branch offices from day one. Candidates who have worked in government before are encouraged to apply. Relevant experience can include technical or policy work in cybersecurity, AI/ML, biosecurity, or China studies, among many other areas. Some fellows will be deeply technical, others more policy-focused, and many will bring a mix of both. Strong candidates across this spectrum should apply.

The program has two tracks: a Fellow track for established professionals (typically 2-6+ years of experience) and a Senior Fellow track for candidates ready to play a senior role in government (typically 10-15+ years of experience). Nearly all roles will be DC-based, but a small number may be performed remotely or from another city where the host agency maintains an office (e.g. San Francisco). US citizenship is required. An existing security clearance, or eligibility to obtain one, is helpful but not required. Salaries range from $170K–$250K+. The application deadline is July 22.

Apply here

Why an AI Rapid Response Fellowship

Advanced AI systems promise exciting gains in science, medicine, and economic growth, among many other areas. But they also present a widening set of security challenges across domains including cyber defense and offense, critical infrastructure protection, and strategic competition with China. Addressing these issues without forfeiting the economic and strategic advantages that increasingly depend on advanced AI is one of the central challenges now facing the federal government.

The release of a new generation of AI models including OpenAI’s GPT 5.5 and Anthropic’s Mythos earlier this spring brought several of these dynamics into sharp public focus. For example, early reporting on Mythos documented meaningful uplift on real-world cyber tasks, with the model surfacing thousands of new high-severity vulnerabilities including issues in every major operating system and web browser. There is still serious debate about the full extent of these models’ capabilities and their implications. But far less contested is the general trajectory: that we are likely to see even more capable models reasonably soon, and that these capabilities will keep diffusing to a wider range of actors.

Policymakers have begun mobilizing in response to these developments. In early June, the White House released two major policy updates: an Executive Order (EO 14409) on advancing AI innovation and security and a companion National Security Presidential Memorandum (NSPM-11) on AI in the national security enterprise. Both the EO and NSPM identify a lack of AI talent in government as a critical bottleneck to delivering on these policy priorities, as have many other observers. As one sponsor of the bipartisan AI Talent Act put it, “The United States can’t fully deliver on its national security mission … if our federal agencies don’t have the talent to meet this moment.”

The AI Rapid Response Fellowship will help close that gap. We are recruiting a specialized cohort of technical and policy talent to work full-time at federal offices on AI security-related issues. The program builds on Horizon’s track record placing technologists in government, with a narrower program scope and a fast-turnaround application timeline suited to the current urgent talent needs in government.

Candidates especially interested in working on AI innovation and science may also be interested in our AI for Science Fellowship, which supports technical leaders in advancing the priorities of the Department of Energy’s Genesis Mission. You can find more information here.

What AI Rapid Response Fellows could focus on

Fellows will support their host offices on some of the most pressing AI questions facing the federal government. These include responding to the cyber capabilities of recent models as well as the wider set of policy and security questions that come with rapidly advancing AI.

The exact portfolio will depend on the office, but the AI EO, NSPM, and statements by senior administration officials have highlighted some of the biggest issues facing the government today. A fellow’s portfolio would likely at least partly overlap with these, which include:

• Facilitating government adoption of AI for defense, intelligence, and national security missions, including the protection of critical infrastructure
• Assessing the capabilities of current AI models and disseminating findings to the public and private actors responsible for security mitigations, such as through vulnerability disclosures
• Protecting frontier model developers from foreign threat actors, for example by hardening critical AI infrastructure and supply chains
• Anticipating and measuring the speed of advancement and diffusion of frontier AI capabilities, including in dual-use capabilities beyond cyber
• Identifying and funding high-priority AI R&D projects, for example on methodologies for evaluating the properties of AI systems and on the technical foundations of AI security, such as robustness and control
• Building out secure AI infrastructure for government, such as high-security AI computing facilities, an AI test range for national security use cases, and shared data and model exchanges across the national security enterprise
• Tracking foreign AI capabilities and their security and economic implications, including intelligence collection and analysis across the AI technology stack

The EO and NSPM mention dozens of agencies and offices that will spearhead relevant efforts, including in the White House, the Intelligence Community, and the Departments of Treasury, War, Energy, Homeland Security, State, and Commerce, among many others.

If you want to dive deeper, Horizon has published an overview of what the AI EO, NSPM, and domain experts say about who’s doing what in government on AI security-related issues and how that may evolve:

Read more about who’s doing what in government on AI security

If accepted to the program, fellows will be introduced to a wide range of offices depending on their interests and background, with offices deciding on fellows’ suitability and portfolios. You can read more about the matching process in the FAQs.

Who should apply

We are looking for technically literate, public service-oriented candidates with the judgment, communication skills, and subject-matter expertise to be effective from day one in a federal policy office.

Core qualifications:

• Expertise relevant to AI security. This could include national security policy, cybersecurity, building AI/ML systems, critical infrastructure protection, biosecurity, US-China policy, intelligence analysis, AI evaluations, supply chain security, or adjacent technical or policy domains. We are open to expertise developed through self-study.
• Relevant work experience. This can include prior policy or government experience focused on AI and/or security issues, as well as technical or leadership work in the private sector—for example, leading projects or teams at an AI startup, security firm, or major technology company.
• Active or prior security clearance preferred. Several potential host offices require clearances, and prior clearance significantly accelerates onboarding. Candidates without clearances are still encouraged to apply, and we expect that not all placements will require a clearance.
• Strong written and verbal communication. Federal policy work depends on the ability to write a clear memo, distill complex technical material for non-technical audiences, and operate effectively in fast-moving interagency settings.
• US citizenship.

If you want to do this kind of work but are unsure whether you meet all the qualifications, we strongly encourage you to apply. Applicants who are not eligible for this program but are excited about entering the field may be redirected to other Horizon opportunities, including the regular Horizon Fellowship program, workshops, or our Career Accelerator. Current government and policy leaders may be eligible for our AI Policy Leadership Network.

You can apply for one or both of the following tracks through the shared application:

• Fellow track. For candidates with technical or policy expertise relevant to AI security—typically 2-6+ years of full-time relevant experience—who are ready to contribute meaningfully in a federal office. We expect most applicants to be considered for this track. Salary range: $170,000–$200,000+, plus an additional stipend for benefits.
• Senior Fellow track. For candidates with substantial prior experience—typically 10-15+ years in relevant technical, policy, or leadership roles—who could play a senior role within a host office. Compensation starts at $250,000, plus an additional stipend for benefits, commensurate with the seniority of the placement.

If you are unsure which track you qualify for, you can indicate interest in both on the application form and we will consider you for the best fit.

If you are unsure whether and where you could serve in government (e.g. based on clearance requirements), we still encourage you to apply. The application and matching process are designed to help you explore your options, and you will not need to make a final commitment to participate until you have offers from host offices to consider. With your consent, applicants may also be referred to other organizations with relevant hiring needs in AI security.

For any questions, email rapidresponse@horizonpublicservice.org.

Program details

Fellows are placed full-time at a federal host office for an initial one-year term. Renewals may be possible depending on funding and host office interest.

In addition to a fully funded placement, fellows receive:

• Connections to leading experts across AI policy and security, and dedicated support in building expertise where needed prior to placement, including briefings on frontier AI capabilities, the federal AI policy landscape, and relevant governmental processes
• Optional participation in the wider Horizon community of AI policy professionals, including relationships with current and former Horizon Fellows and invitations to happy hours, workshops, and topical convenings
• Continued access to career support and the Horizon network after the fellowship concludes, including connections to people and organizations working on AI security-related issues across government, think tanks, industry, and academia
• A professional development budget for conferences, training, and related career development resources and opportunities, plus a relocation stipend and support for candidates moving to the Washington, DC area

Timeline:

• Application process. The deadline to apply is July 22. Applicants who advance will be assessed based on a combination of interviews, a short work test, and reference checks in the subsequent weeks.
• Selection and offers. Initial offers will go out to finalists by early-mid September, and responses are expected within a week. Accepting an offer commits a finalist only to exploring matches with prospective host offices—not to any particular office, and not to ultimately participating in the program. Finalists don’t need to make a final decision about joining the program until they have a concrete offer from a host office.
• Host office matching. As soon as possible after a finalist is selected, Horizon will begin introducing them to prospective host offices in government. Finalists may explore multiple offices to find the best fit, with Horizon facilitating conversations throughout.
• Earliest start date. Start date is determined in consultation with the fellow and host office, and contingent on administrative requirements (e.g. hiring authority, security clearance), but is generally expected to be as soon as possible. (If you are matched with a federal agency that has a multi-month onboarding process, you may be able to join a nonprofit organization doing relevant work in preparation for your placement while administrative processing is pending.)

Apply here

Frequently Asked Questions

If your question is not answered here, please email rapidresponse@horizonpublicservice.org.

The regular Horizon Fellowship helps people transition into careers in technology policy. It provides training to prepare fellows to serve in government and places them for one to two years across think tanks, Congress, and the executive branch. The AI Rapid Response Fellowship is for people who are ready to serve in government from day one with minimal training. It places fellows only in the executive branch, and on a faster application timeline built to get people into government as soon as possible. While the standard Fellowship is aimed at people pivoting into a policy career more broadly, the Rapid Response Fellowship also suits people who have worked in government before and people who want a short tour-of-duty in government to respond to this moment. Salaries and seniority skew higher than the standard fellowship program, and the initial placement is one year with no default renewal.

Yes. They’re separate applications, and applying to one doesn’t affect your candidacy for the other.

No. A current or prior clearance is preferred, since it broadens the set of host offices you can place into and may accelerate how quickly you can start a placement, but candidates without clearances are still encouraged to apply. Some host offices can accommodate clearance processing during the placement, and some will not require a clearance.

Yes. AI raises issues across a wide range of fields, and the government needs expertise that extends well beyond the areas we were able to name. The list in the description is illustrative, not exhaustive. If you think your background could bear on how the government responds to, deploys, and secures the latest generations of AI systems, we encourage you to apply.

Yes. The needs of federal agencies vary widely in how technical or policy-focused they are, and strong candidates will sit at both ends. Some fellows will be deeply technical, like a vulnerability researcher who runs frontier models against hard targets and confirms which of the flaws they surface are real. Others might be policy specialists who can translate between the tech world and the interagency process, making sure policy conversations are informed by a broadly accurate understanding of AI concepts and capabilities. And others will fall somewhere in between.

Yes. We are aiming to begin placements as soon as candidates and host offices are ready, with the earliest starts likely this fall. Start dates are set in consultation with you and your host office, and there is some room for flexibility, including for candidates who need time to wrap up current commitments or to complete administrative steps like a clearance. If timing is your main hesitation, apply and we can work through the options.

These are full-time placements. For the duration of the fellowship you would step away from your current job and work within your host office; for most participants, we expect this would require leaving your current employer. If your current employer allows for a public service sabbatical, this may be compatible with program participation pending appropriate conflict of interest mitigations. Please indicate if you intend to use such an opportunity in your application.

Yes. The application and matching process are built to help you explore what is possible before you commit to anything. Both AI itself and AI policy are changing quickly, and many new events and opportunities will likely come up in the coming months. Applying lets you see which roles exist and talk with prospective host offices before deciding whether any of them is a great fit. You will not need to commit to joining until you have a concrete offer in hand.

Yes. If you already work in government and have relevant expertise, this program can help you move into a role focused more directly on AI security.

Specific placements depend on host office openings, candidate fit, and clearance status, and are confirmed during the matching phase of the program. We work closely with both candidates and host offices to identify strong matches. If you want to get a feel for the kinds of offices and issues involved, our overview of who’s doing what in government on AI security maps the landscape in more detail.

Once you are selected as a finalist, Horizon begins introducing you to prospective host offices based on your background, interests, and clearance status. You will typically speak with several offices to understand the role and the team and to gauge fit on both sides. You can explore more than one office at the same time. Matching will run on a rolling basis, so the timeline varies from one candidate to the next.

Yes. The process is collaborative. The application asks about your interests, you speak with prospective offices directly, and no placement is finalized until you have an offer you want to accept. You will not be placed anywhere you have not agreed to.

No. Accepting a place in the program commits you only to exploring matches with prospective host offices, not to joining any particular one. You would not need to leave your current job until you have a concrete offer from a host office that you have decided to accept.

If your host office has a long or uncertain onboarding process, Horizon can fund an interim placement at a nonprofit doing relevant AI security work, so you can start contributing and building context while your processing is pending. You could also stay in your current role until the placement is ready. If a security clearance ultimately does not come through, not every host office requires one, so we can explore an alternative placement.

If you want to stay in government, you may be able to move into a permanent role, either at your host office or another office. Transitional support beyond your year-long fellowship may be available from Horizon, contingent on funding. If you would rather move on to something else, you are welcome to do so; there is no expectation beyond the one-year placement. Either way, you keep access to the Horizon network and career support after the fellowship ends.

It depends on the office and the role. Most federal positions involve a suitability or fitness determination, and some also require a background investigation or security clearance depending on the level of access involved. A smaller number of roles are political appointments that go through the appointing office’s own process. Requirements vary with the role and where it sits, so during matching we can give you a clearer picture of what to expect for the specific offices you’re exploring.

It depends on the office and the role, and it is something we help you think through before you commit. Many federal positions require financial disclosure, either confidential or public. Federal conflict-of-interest rules can also require you to recuse yourself from specific matters or, in some cases, to divest particular holdings. This comes up most often when someone holds equity or options in companies their work would touch directly, such as AI developers or government contractors. We will flag the likely implications for the offices you are exploring so you can weigh them in advance.

About Horizon

The Horizon Institute for Public Service is a nonpartisan nonprofit that helps the US government navigate rapid technological change by building the next generation of emerging technology policy talent. Horizon supports people across a range of career stages—through fellowship placements, a career accelerator, workshops, senior policy networks, and online career resources.

The Horizon Fellowship, our flagship program, places fellows at host organizations to work on policy related to artificial intelligence, biotechnology, and other emerging technologies. This Rapid Response Fellowship builds on that model with a focused effort to help government respond to urgent AI security needs. Learn more about Horizon and our programs here.