Anthropic Fellows Program, AI Safety

<div class="content-intro"><h2><strong>About Anthropic</strong></h2> <p>Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems.</p></div><p><a href="https://bit.ly/afpsafety"><strong>Apply using this link</strong></a><strong>. </strong>Applications for the next cohort of Anthropic Fellows close at <strong>11:59pm PT on July 26</strong>. The cohort is expected to start <strong>November 2</strong>. In some circumstances, we can accommodate fellows starting outside the usual cohort timelines — please note in your application if the November start date doesn't work for you.</p> <p>This page is specific to one of the Anthropic Fellows Workstreams, see also the main&nbsp;<strong><a href="https://job-boards.greenhouse.io/anthropic/jobs/5023394008">Anthropic Fellows posting</a></strong>.</p> <h1><strong>Anthropic Fellows Program overview</strong></h1> <p>The Anthropic Fellows Program is designed to foster AI research and engineering talent. We provide funding and mentorship to promising technical talent - regardless of previous experience.</p> <p>Fellows will primarily use external infrastructure (e.g. open-source models, public APIs) to work on an empirical project aligned with our research priorities, with the goal of producing a <strong>public output</strong> (e.g. a paper submission). In one of our earlier cohorts, over 80% of fellows produced papers.&nbsp;We run multiple cohorts of Fellows each year and review applications on a rolling basis.</p> <h2><strong>What to expect</strong></h2> <ul> <li>4 months of full-time research&nbsp;</li> <li>Direct mentorship from Anthropic researchers&nbsp;</li> <li>Access to a shared workspace (in either Berkeley, California or London, UK)</li> <li>Connection to the broader AI safety and security research community</li> <li>Weekly stipend of 3,850 USD / 2,310 GBP / 4,300 CAD + benefits (these vary by country)</li> <li>Funding for compute (~$15k/month) and other research expenses</li> </ul> <h2><strong>Interview process</strong></h2> <p>The interview process will include an initial application &amp; reference check, technical assessments &amp; interviews, and a research discussion.&nbsp;</p> <p><strong>We encourage you to apply even if you do not believe you meet every single qualification.</strong> Not all strong candidates will meet every single qualification as listed. Research shows that people who identify as being from underrepresented groups are more prone to experiencing imposter syndrome and doubting the strength of their candidacy, so we urge you not to exclude yourself prematurely and to submit an application if you're interested in this work. We think AI systems like the ones we're building have enormous social and ethical implications. We think this makes representation even more important, and we strive to include a range of diverse perspectives on our team.</p> <h2><strong>Compensation</strong></h2> <p>The expected base stipend for this role is 3,850 USD / 2,310 GBP / 4,300 CAD per week, with an expectation of 40 hours per week for 4 months (with possible extension).</p> <h2><strong>Fellows workstreams</strong></h2> <p>Due to the success of the <a href="https://alignment.anthropic.com/2024/anthropic-fellows-program/">Anthropic Fellows for AI Safety Research</a> program, we are now expanding it across teams at Anthropic. We expect there to be significant overlap in the types of skills and responsibilities across the roles and will by default consider candidates for all the workstreams.</p> <p>Some of the workstreams may include unique assessment steps; we therefore ask you for workstream preferences in the<a href="https://bit.ly/afpsafety"> application</a>. You can see an overview of the current workstreams below:</p> <ol> <li>AI Safety Fellows</li> <li>AI Security Fellows</li> <li>ML Systems &amp; Performance Fellows</li> <li>Reinforcement Learning Fellows</li> <li>Economics &amp; Societal Impacts Fellows</li> </ol> <p>This page is specific to one of the Anthropic Fellows Workstreams, see also the main&nbsp;<strong><a href="https://job-boards.greenhouse.io/anthropic/jobs/5023394008">Anthropic Fellows posting</a></strong>.</p> <h2><strong>Across the workstreams, you may be a good fit if you:</strong></h2> <ul> <li>Are motivated by making sure AI is safe and beneficial for society as a whole</li> <li>Are excited to transition into empirical AI research and would be interested in a full-time role at Anthropic</li> <li>Have a strong technical background in computer science, mathematics, or physics</li> <li>Thrive in fast-paced, collaborative environments</li> <li>Can implement ideas quickly and communicate clearly</li> </ul> <h2><strong>Strong candidates may also have:</strong></h2> <ul> <li>Strong background in a discipline relevant to a specific Fellows workstream (e.g. economics, social sciences, or cybersecurity)</li> <li>Experience in areas of research or engineering related to their workstream</li> </ul> <h2><strong>Candidates must be:</strong></h2> <ul> <li>Fluent in Python programming</li> <li>Available to work full-time on the Fellows program</li> </ul> <h1><strong>AI Safety Fellows</strong></h1> <h2><strong>Mentors, research areas, &amp; past projects</strong></h2> <p>Fellows will undergo a project selection &amp; mentor matching process. Potential mentors include:</p> <ul> <li>Sam Bowman</li> <li>Sara Price</li> <li>Alex Tamkin</li> <li>Nina Panickssery</li> <li>Trenton Bricken</li> <li>Logan Graham</li> <li>Jascha Sohl-Dickstein</li> <li>Joe Benton</li> <li>Collin Burns</li> <li>Fabien Roger</li> <li>Samuel Marks</li> <li>Kyle Fish</li> <li>Ethan Perez</li> </ul> <p>Our mentors will lead projects in select AI safety research areas, such as:</p> <ul> <li><strong>Scalable Oversight:</strong> Developing techniques to keep highly capable models helpful and honest, even as they surpass human-level intelligence in various domains.</li> <li><strong>Adversarial Robustness and AI Control:</strong> Creating methods to ensure advanced AI systems remain safe and harmless in unfamiliar or adversarial scenarios.</li> <li><strong>Model Organisms:</strong> Creating model organisms of misalignment to improve our empirical understanding of how alignment failures might arise.</li> <li><strong>Model Internals / Mechanistic Interpretability:</strong> Advancing our understanding of the internal workings of large language models to enable more targeted interventions and safety measures.</li> <li><strong>AI Welfare:</strong> Improving our understanding of potential AI welfare and developing related evaluations and mitigations.</li> </ul> <p>On our <a href="https://alignment.anthropic.com/">Alignment Science</a> and <a href="https://red.anthropic.com/">Frontier Red Team</a> blogs, you can read about past projects, including:</p> <ul> <li><a href="https://alignment.anthropic.com/2025/subliminal-learning/">Subliminal Learning: Language Models Transmit Behavioral Traits via Hidden Signals in Data:</a> Alex Cloud and Minh Le, et al., mentors including Samuel Marks and Owain Evans</li> <li><a href="https://www.anthropic.com/research/open-source-circuit-tracing">Open-source circuits:</a> Michael Hanna and Mateusz Piotrowski with mentorship from Emmanuel Ameisen and Jack Lindsey</li> </ul> <p>For a full list of representative projects for each area, please see these blog posts: <a href="https://alignment.anthropic.com/2024/anthropic-fellows-program/#:~:text=We're%20launching%20the%20Anthropic,time%20on%20AI%20safety%20research.">Introducing the Anthropic Fellows Program for AI Safety Research</a>, <a href="https://alignment.anthropic.com/2025/recommended-directions/">Recommendations for Technical AI Safety Research Directions</a>.</p> <h2><strong>Unique candidate cr