Wiktoria Kozyra
United Kingdom
PhD Student
University of Oxford & University of Cambridge
Neuroscience/Cognitive Science
United Kingdom PhD Student University of Oxford & University of Cambridge Neuroscience/Cognitive Science
Project
Synthetic pleasure: Can a robot have an orgasm? Liking processing beyond reward and reinforcement learning.
Project Overview
Reinforcement-learning proposes how biological and artificial systems update reward predictions and adapt their behaviour, but not why biological reward is accompanied by the subjective sense of pleasure. Artificial RL algorithms demonstrate that pleasure is not necessary for a successful reward-maximising behaviour. What is more, similar mechanisms seem to be linked to dopamine-based reward processing. The question remains why biological organisms experience pleasure if all the computational aspects of reward maximisation can be implemented unconsciously. This project asks: what functions does pleasure serve and what features would make an artificial system a candidate for experiencing pleasure? We will conduct an integrative review spanning affective neuroscience, comparative cognition, evolutionary biology, reinforcement learning and consciousness science. It will look at the difference between the explanatory scope of dopamine, particularly reward prediction error, and the hedonic liking involving opioid and endocannabinoid systems, homeostatic regulation, and representations of subjective value. Evidence across animals will be assessed cautiously, including behavioural and neural markers of hedonic processing. The principal outcome will be an opinion article proposing candidate functions of pleasure and provisional implications for artificial agents. Human experiments and computational models will be formulated possibly as further extensions.
Mentor
About the Scholar
I am a DPhil student at the University of Oxford, where I work with Prof. Matthew Rushworth and Dr Jan Grohn on cognitive computational neuroscience, focusing on how people infer generative models and latent causes. During my MSc in the same lab, I used fMRI to study distributional reinforcement learning in human reward processing. I am currently also completing an MPhil at the University of Cambridge with Dr Andrea Luppi and Prof. Emmanuel Stamatakis, working on macaque neural correlates of consciousness, particularly functional-connectivity changes during anaesthesia-induced loss of consciousness and deep-brain-stimulation-mediated recovery. Before this, I completed my BSc at UCL, where I worked with Prof. Steve Fleming on Bayesian models of metacognition in humans and convolutional neural networks, and with Prof. Patrick Haggard on EEG signatures of voluntary movement generation. I am originally from Poland, where I previously studied Mathematics and Economics at the University of Warsaw.
Why AISS?
"Consciousness is my main scientific and philosophical interest, and the motivation behind all my work. For my project with Axel, I am especially excited to think about what distinguishes systems that learn to behaviourally maximise rewards from those that have the capacity to experience pleasure. At a time when AI systems are becoming increasingly behaviourally human-like, I believe rigorous research on consciousness and sentience is more important than ever, especially because our intuitive judgements about consciousness attribution can be powerful but misleading. I am very grateful that this much-needed programme has been created, and honoured to be part of the AI Sentience Scholars cohort."
Links
Updates
July 2026 - AI Sentience Mentor Axel Cleeremans and AI Sentience Scholar Wiktoria Kozyra meet up at the Association for the Scientific Study of Consciousness conference in Santiago, Chile! Learn more.