The rapid advancement of artificial intelligence is creating unprecedented opportunities for researchers, engineers, economists, policy professionals, and technical experts worldwide. As governments, industries, and institutions increasingly rely on advanced AI systems, the demand for professionals capable of building safe, interpretable, and beneficial AI technologies continues to grow.
To support the next generation of frontier AI talent, Anthropic has officially opened applications for the highly competitive Anthropic Fellows Program 2026 — a fully funded, four-month research fellowship focused on artificial intelligence safety, security, reinforcement learning, governance, and advanced machine learning systems.
The fellowship is designed to help talented individuals transition into high-impact AI research careers by providing mentorship, financial support, compute resources, and opportunities to collaborate directly with leading AI researchers working on some of the world’s most important AI safety challenges.
Applications are accepted on a rolling basis for cohorts beginning in late September 2026 and beyond.
About the Anthropic Fellows Program 2026
The Anthropic Fellows Program is a full-time research initiative that provides participants with the opportunity to conduct empirical AI research projects aligned with Anthropic’s frontier research priorities.
Fellows work under the guidance of experienced researchers, engineers, and policy experts while contributing to projects that advance safe and beneficial artificial intelligence systems.
The programme is specifically designed for individuals who may already possess strong technical or analytical abilities but want structured mentorship, research experience, and professional exposure within the rapidly growing AI ecosystem.
According to Anthropic, more than 80% of fellows from earlier cohorts successfully produced public research outputs, including papers, evaluations, technical analyses, and open-source contributions.
What Fellows Will Work On
Participants in the fellowship may join different workstreams depending on their background, interests, and expertise.
Available Fellowship Workstreams
- AI Safety Fellows
- AI Security Fellows
- Machine Learning Systems and Performance Fellows
- Reinforcement Learning Fellows
- Economics and Policy Fellows
Each workstream focuses on solving critical frontier AI problems and advancing the global understanding of safe AI deployment.
Key Research Areas
Fellows may contribute to research topics including:
- Scalable oversight systems for advanced AI
- AI alignment and control techniques
- Adversarial robustness and AI security
- Mechanistic interpretability
- AI welfare and ethics
- Reinforcement learning systems
- AI-enabled cybersecurity risks
- AI governance and public policy
- Economic impacts of AI systems
- Large language model evaluation
- Safety testing and red teaming
- Distributed AI systems and infrastructure
- AI model training and optimization
Participants will collaborate with mentors who are actively shaping the global AI safety landscape.
Fellowship Benefits
The Anthropic Fellows Program provides substantial financial, technical, and professional support to selected participants.
Financial Support
Selected fellows receive weekly stipends based on their location:
- 3,850 USD per week
- 2,310 GBP per week
- 4,300 CAD per week
The fellowship runs full-time for four months, with possible extensions in some cases.
Additional Benefits
Participants also receive:
- Direct mentorship from Anthropic researchers
- Access to shared workspaces in:
- London, United Kingdom
- Berkeley, California
- Remote participation opportunities
- Funding for compute resources
- Approximately 15,000 USD monthly compute budget
- Research expense support
- Networking opportunities within the global AI safety ecosystem
- Exposure to cutting-edge AI research projects
- Potential pathways into full-time AI careers
Who Should Apply?
Anthropic encourages applications from individuals with diverse academic, technical, and professional backgrounds.
Strong candidates may come from fields such as:
- Computer Science
- Artificial Intelligence
- Machine Learning
- Mathematics
- Physics
- Economics
- Cybersecurity
- Public Policy
- Engineering
- Data Science
- Social Sciences
- Software Development
The organization explicitly encourages applications from underrepresented groups and individuals transitioning into AI research and governance careers.
Even candidates who do not meet every listed qualification are encouraged to apply.
Eligibility Requirements
To qualify for the fellowship, applicants must generally:
- Be fluent in Python programming
- Be available full-time for four months
- Possess strong analytical and communication skills
- Demonstrate interest in safe and beneficial AI systems
- Be capable of collaborative research work
Work Authorization Requirement
Applicants must already possess work authorization in one of the following countries:
- United States
- United Kingdom
- Canada
Visa sponsorship is currently unavailable for fellows.
AI Safety Fellows Track
The AI Safety Fellows track focuses on reducing catastrophic risks from advanced AI systems.
Research Topics Include
- Scalable oversight
- Adversarial robustness
- AI control methods
- Model organisms of misalignment
- Mechanistic interpretability
- AI welfare evaluations
Ideal Candidates May Have
- Experience with empirical machine learning research
- Familiarity with large language models
- Open-source research contributions
- Interest in AI alignment and catastrophic risk reduction
AI Security Fellows Track
The AI Security Fellows track explores AI-enabled cyber risks, vulnerability analysis, and adversarial systems.
Research Areas Include
- AI red teaming
- Offensive security research
- Vulnerability discovery
- Security evaluation systems
- AI-assisted exploit analysis
Strong Candidates Often Have
- Pentesting experience
- Vulnerability research experience
- Open-source security contributions
- Experience with ML experimentation
- Deep learning systems knowledge
Machine Learning Systems and Performance Fellows
This workstream focuses on large-scale AI infrastructure, distributed systems, and performance optimization.
Potential Projects Include
- Accelerator infrastructure
- Synthetic data pipelines
- High-performance distributed systems
- ML systems engineering
- Model evaluation pipelines
Reinforcement Learning Fellows
Participants in this track focus on reinforcement learning systems, training environments, and evaluation methods.
Areas of Focus Include
- RL algorithms
- Safety-focused RL environments
- AI training data quality
- Generalization research
- Large language model evaluation
Economics and Policy Fellows
The Economics and Policy track supports research examining the societal and economic impacts of advanced AI systems.
Possible Research Topics Include
- AI and labor markets
- AI governance frameworks
- Public policy recommendations
- Societal resilience to AI risks
- Economic forecasting for AI systems
- Cyber and biosecurity implications of advanced AI
This track is ideal for individuals interested in AI governance, economic policy, and interdisciplinary AI impact research.
Programme Logistics
Duration
- Four months full-time
Work Locations
Participants may work:
- Remotely from eligible countries
- From Berkeley, California
- From London, United Kingdom
Remote Participation
Remote participation is available only for applicants based in:
- United States
- United Kingdom
- Canada
Application Process
The selection process includes several stages:
- Initial application review
- Reference checks
- Technical assessments
- Research interviews
- Research discussion rounds
Applications are reviewed on a rolling basis, meaning early applications are strongly encouraged.
Career Outcomes and Opportunities
The Anthropic Fellows Program serves as a major talent pipeline into AI safety and governance careers.
Participants gain:
- Research experience
- Professional mentorship
- Publication opportunities
- Industry exposure
- Networking opportunities
- Access to leading AI researchers
Historically, between 25% and 50% of fellows received full-time offers after completing the programme.
Even fellows who do not join Anthropic often transition into influential roles across AI safety organizations, research labs, governance institutions, and technical policy teams.
Why This Fellowship Matters
As artificial intelligence systems become increasingly powerful, the need for professionals capable of ensuring these technologies remain safe, interpretable, and beneficial has never been greater.
The Anthropic Fellows Program stands out because it combines:
- Frontier AI research
- Practical mentorship
- Financial support
- Public-impact research
- Career acceleration
- AI governance exposure
For aspiring AI researchers, engineers, economists, and policy professionals, the programme offers a rare opportunity to contribute directly to some of the most important AI safety challenges facing society today.
How to Apply
Interested applicants can learn more and submit applications through the official Anthropic careers page.
Applications are accepted on a rolling basis for future cohorts beginning in late September 2026 and beyond.
For more information about this opportunity, visit here.
Discover more Global Fellowship Opportunities on OFY here.
