Ananya is an interdisciplinary researcher working at the intersection of AI safety, human-AI interaction and behavioural evaluation. Her background is in anthropology and history (MPhil, University of Cambridge; BA First Class Honours, Goldsmiths, University of London) and she holds a Churchill Fellowship for international research on accessibility as a socio-technical system.

Her research focuses on what she terms the relational layer of AI: the space where humans and AI systems shape each other in ways that current safety frameworks are not designed to detect. Her founding pilot study demonstrated that user emotional framing systematically changes how AI models communicate risk, hold boundaries and validate decisions, with effects that are invisible to standard output-level evaluations but significant for safety in high-stakes contexts.

Before entering AI safety, Ananya spent over eight years designing rigorous qualitative research for complex socio-technical environments. She led research projects for global organisations including Sanofi, L'Oréal and Colgate at C Space, managed C-suite client relationships at Gartner and consulted on campaign strategy for Greenpeace UK, where her work contributed directly to policy change in UK Parliament. She has spoken at institutions including the Tate Modern, and worked with organisations spanning the United Nations, YouTube, the Wellcome Trust, the BBC and the Guardian.

Her path to founding the Institute reflects a consistent thread across her career: the study of how systems succeed or fail at the human level, and the conviction that the most consequential dynamics are often the ones that existing frameworks are not designed to see. In AI safety, that conviction led her to ask a question that the field had not yet formalised: if AI-human interactions are relationships, what does relationship science tell us about how to evaluate them?

The Institute of Relational AI is her answer.