← All mentors

Mentor · Summer 2027

Pranav Mahajan

Postdoctoral Researcher, University of Oxford

Cognitive sciencePsychologyComputational neuroscience

About

Pranav is a postdoctoral researcher at the University of Oxford working across computational cognitive neuroscience, health AI, and AI safety. Pranav studies the adaptive and maladaptive sides of human and artificial minds, with the aim of helping build AI systems that work for people. Pranav’s PhD at Oxford was on safe learning in humans and machines, contributing to safe reinforcement learning with applications to safer robotics and to health technologies for anxiety disorders and chronic pain. Before the PhD, Pranav worked on scalable distributed deep learning at Nvidia.

Research directions

  • Cognitive mechanisms by which AI users can develop co-dependency or experience gradual disempowerment in the emotional domain through reliance on AI
  • Whether AIs can be built to prioritise users’ long-term well-being, through a simple change to post-training or a radically different approach that balances short-term and long-term rewards
  • Using ideas from cognitive science to improve the science of evaluation, user modelling, and multi-LLM and multi-agent dynamics
  • Applying LLM evaluations to healthcare (mental health, chronic pain) and studying the societal resilience aspects of AI uptake

Recent work

  • Mind the Gap: How Elicitation Protocols Shape the Stated-Revealed Preference Gap in Language Models (Eval Eval at ACL 2026; TAIS 2026)

Profile last updated:

More mentors