Anthropic is inviting applications for its Conceptual Reasoning Fellow programme, a four-month research fellowship for experts in philosophy, cognitive science, economics, public policy and law who are interested in advancing the reasoning capabilities of artificial intelligence. The fellowship brings experienced researchers and subject-matter experts together with Anthropic’s research teams to investigate complex questions that require expert judgment. Fellows will work with Claude, Anthropic’s advanced AI model, to evaluate its capabilities, identify weaknesses and develop methods for improving its performance in specialised fields. The programme is scheduled to run from 12 January to 8 May 2027. Applications close on 28 October 2026, with offers expected to be made by 20 November 2026. Successful candidates will receive compensation ranging from US$100,000 to US$167,000 for the four-month fellowship, depending on the applicable arrangement, alongside other benefits. About the Conceptual Reasoning Fellowship Anthropic develops AI systems intended to be reliable, interpretable and steerable, with the broader goal of ensuring that artificial intelligence benefits society. The Conceptual Reasoning Fellowship focuses on questions that cannot always be answered through straightforward factual verification. Complex legal, philosophical, economic and policy problems often require contextual understanding, critical thinking and nuanced professional judgment. Fellows will collaborate with Anthropic researchers to explore how frontier AI systems can support these forms of reasoning. Each participant will undertake a defined research project matched to their background, supported by mentorship, access to Claude models and internal evaluation tools, and collaboration with other fellows. The programme offers an opportunity for experienced professionals to contribute their disciplinary expertise to the development and evaluation of advanced AI systems. Key Responsibilities Selected fellows will contribute to research activities that assess and improve Claude’s performance in their respective fields. Responsibilities include: Identifying important unanswered questions in their disciplines and evaluating approaches to addressing them using Claude. Investigating major societal challenges that require specialised knowledge and assessing where AI systems perform well or poorly. Designing evaluations, benchmarks and reinforcement-learning environments that systematically test model reasoning. Translating expert knowledge into practical tests that measure the quality, reliability and limitations of AI-generated analysis. Collaborating with other fellows and Anthropic researchers to develop challenging tasks requiring advanced judgment. Illustrative Research Projects The fellowship welcomes proposals across several disciplines. Potential projects include: Philosophy: Developing difficult ethical problems or formal epistemology tasks that test whether AI models can reason through competing principles and complex judgments. Cognitive science: Designing experimental tasks that assess model comparison, hypothesis testing and research-design decisions. Economics: Creating economic modelling and identification problems with reliable ground-truth answers to evaluate analytical accuracy. Public policy: Developing regulatory analysis and cost-benefit assessment tasks with expert-defined evaluation criteria. Law: Building datasets relevant to legal reasoning and AI policy, allowing researchers to assess how effectively models interpret and analyse legal questions. These examples illustrate possible directions rather than fixed project assignments. Fellows will work with mentors to refine their research questions and project scope. Eligibility Requirements Applicants must meet the following minimum requirements: Hold a PhD, be completing a PhD in philosophy, cognitive science, economics, public policy or law, or demonstrate equivalent research experience. Have practical experience using Claude and Claude Code in their regular workflow. Be enthusiastic about the potential for AI to reason through difficult questions in their area of expertise. Be comfortable working in a fast-paced environment with substantial independence. Be able to deliver concrete research outcomes within the four-month programme. Have existing authorisation to work in the United States for the entire fellowship period. Preferred qualifications include experience designing evaluations, benchmarks, datasets or expert assessment rubrics. Demonstrated disciplinary expertise through research, scholarship or high-level analytical writing is also advantageous. Fellowship Duration, Location and Compensation The fellowship runs full-time from 12 January to 8 May 2027. Some placements may be considered for an extension. The default location is San Francisco, California. Anthropic may consider remote arrangements within the United States in certain circumstances. All fellows must attend a one-week in-person orientation in San Francisco. Compensation ranges from US$100,000 to US$167,000 for the four-month period, equivalent to an annualised rate of approximately US$300,000–US$500,000. Other benefits are also provided. The fellowship is a fixed-term contract, with participants engaged through an employer of record rather than directly employed by Anthropic. Important: Anthropic is unable to sponsor visas for this fellowship. Applicants must already have, or independently obtain, full-time US work authorisation. International applicants without the required authorisation will not meet this condition. Application Deadline and Selection Process Applications close on 28 October 2026. Anthropic plans to make offers by 20 November 2026. Later selection stages will typically include a take-home exercise and discussions with researchers and potential mentors about applicants’ research interests and possible projects. Applicants should prepare a current CV or résumé, provide a LinkedIn profile or résumé as required by the form, identify two professional references and explain their proposed research ideas. The application also requests information about Claude Code experience, previous AI or machine-learning work, and the disciplinary field in which the candidate intends to contribute. How to Apply Interested candidates should visit Anthropic’s official careers website, locate the Conceptual Reasoning Fellow position and complete the online application before the deadline. VISIT OFFICIAL WEBSITE TO APPLY Applicants should confirm the current vacancy and submit their information through Anthropic’s official recruitment platform. Genuine Anthropic recruiters use official @anthropic.com email addresses, and applicants should never pay recruitment fees. For more opportunities such as these please follow us on Facebook, Instagram, Twitter, LinkedIn and WPChannel Disclaimer: Global South Opportunities (GSO) does not take ownership or credit for the opportunities shared on this platform. All opportunities are sourced from official announcements and external sources for informational purposes only. Applicants are encouraged to verify all details, deadlines, and eligibility requirements from the official source before applying. Opportunities may be updated or changed by the providing organisations.