Safe AI Forum
We work to advance technically grounded international coordination to reduce catastrophic risks from frontier AI, with a particular focus on China and Western countries. We translate technical AI safety and governance tools into practical proposals for cooperation.
Stream overview
We work on technically grounded international coordination to reduce risks from frontier AI, with a particular focus on China-West relations. Fellows may contribute to projects such as:
- Joint testing and evaluation interoperability: Developing proposals for how Chinese and Western institutions could align or mutually recognise frontier AI safety testing and evaluation practices. This may include comparing risk thresholds, evaluation methodologies, benchmark design, reporting formats, and procedures for securely sharing relevant safety evidence.
- AI incidents and crisis management: Researching mechanisms for preventing, reporting, and responding to serious frontier-AI incidents, including misuse, unexpected model behaviour, or failures of safety controls. Fellows may develop proposals for incident-taxonomy and reporting standards, communication channels, joint investigation procedures, and crisis-management protocols.
- Technical governance for international cooperation: Assessing how tools such as capability and risk evaluations, compute monitoring, safety cases, model-release safeguards, and verification mechanisms could support credible international commitments. This could also include analysing the geopolitical implications of advanced AI and identifying practical areas for China-West coordination.
Mentors
Isabella Duan is a senior researcher at Safe AI Forum. She previously interned at Google DeepMind and the Centre for the Governance of AI, developing social impact evaluations for frontier AI models.
Yawen Duan is a Senior Researcher at the Safe AI Forum (SAIF), where he works on international coordination and technical AI governance for managing extreme risks from advanced AI. His research focuses on risks from increasingly agentic AI systems: how to set and operationalize risk thresholds and red lines, and how to evaluate and monitor agents post-deployment. Previously, as AI Safety Research Manager at Concordia AI, he led the Frontier AI Risk Management Framework (co-published with Shanghai AI Lab).
Mentorship style
1 hour weekly meetings by default for high-level guidance. We are active on Slack and typically respond within a day for quick questions.
Fellows we are looking for
- A strong working understanding of frontier AI governance and safety, including major international initiatives and institutions such as the AI Safety Institute network, AI Safety Summits, and relevant multilateral processes.
- Familiarity with Chinese AI governance and the Chinese AI ecosystem, including key institutions, policy developments, safety debates, and relevant technical or regulatory trends.
- Familiarity with at least one major frontier AI risk domain, such as AI-enabled cyber capabilities, CBRN risks, loss of control, or systemic risks from highly capable AI, as well as the role of evaluations and mitigations in that domain.
- A working understanding of US-China relations and technology competition, including how strategic dynamics may shape incentives and constraints for cooperation on advanced AI.
- A demonstrated commitment to reducing catastrophic risks from frontier AI and to engaging constructively across Chinese and Western perspectives. Applicants should be motivated to engage seriously and constructively with both Chinese and Western perspectives, including where those perspectives differ.
Project selection
We will provide a shortlist of projects that we are keen for the scholar to work on in Week 1. We'll ask scholars to scope these in the 1st week and make a determination about which project to focus on in Week 2.
The Winter 2026 cohort offers a wide range of research streams led by experts across AI alignment, interpretability, governance, and safety. Each stream provides its own research agenda, methodology, and mentorship focus.
