Sarah Ball

Research Interests

• AI Safety
• AI Alignment
• Mechanistic Interpretability
• Social Impact of AI
• Human-AI Cooperation

Short Description

Sarah is a PhD student committed to ensuring generative AI systems are safe and aligned with human values. In her work, she investigates how adversarial attacks circumvent model safety mechanisms and explores the boundaries of AI’s ability to replace human judgement. Her research is supported by the Konrad Zuse School of Excellence in Reliable AI and the Munich Center for Machine Learning.

During fall semester 2024, Sarah was a visiting research fellow at the Simons Institute for the Theory of Computing at the University of California, Berkeley, where she collaborated with leading researchers and connected with experts from organizations including OpenAI and Anthropic.

Before joining the lab at LMU, Sarah was a Master student in Social Data Science at the University of Oxford where she graduated top of her cohort, won the department’s thesis prize, and got selected by the university as a student prize winner to be recognized at Encaenia – Oxford University’s flagship ceremony to reward honorary degrees.

Beyond her research, Sarah actively contributes to public discourse on AI by communicating AI research to policymakers, industry leaders, and the general public through TV interviews, talks, podcasts, and panel discussions.