Research

Practice-oriented AI is about bridging the gap between between complex problem domains such as those found in science and research, and AI algorithms and techniques that could be used to solve problems in those domains.

Within the broad area of practice-oriented AI, my current research interests focus on:

Research Interest

My research lies at the intersection of Artificial Intelligence, cognitive science, behavioural economics, and decision science. I investigate how human reasoning shapes the behaviour of AI systems and how AI can be designed to support better decision-making in high-stakes environments, particularly healthcare.

A central premise of my research is that AI systems learn from data generated through human decisions. Because human decision-making often relies on heuristics, efficient mental shortcuts that simplify complex problems, AI models can inherit and reinforce the cognitive biases embedded in those decisions. Understanding how these biases emerge, propagate through data, and influence AI behaviour is fundamental to developing trustworthy and reliable intelligent systems.

My research seeks to understand not only how AI inherits human reasoning patterns but also how AI can actively support human decision-making. By combining insights from artificial intelligence, cognitive science, and behavioural economics, I aim to develop AI systems that are both technically robust and cognitively informed.

Research Themes

My current research is organised around four interconnected themes.

Characterising Human Reasoning

I investigate how heuristic reasoning and cognitive biases influence human decision-making, particularly in clinical environments. This work aims to understand how these reasoning patterns become reflected in the data used to develop AI systems.

Detecting and Quantifying Cognitive Biases

I develop computational approaches to identify, characterise, and quantify heuristic-driven behaviours in both human decision-making and AI models trained on human-generated data. Making these patterns observable is an important step towards understanding their impact on AI systems.

Bias-Aware and Trustworthy AI

I study methods for developing AI systems that recognise and mitigate inherited cognitive biases while maintaining predictive performance, interpretability, fairness, and robustness. My goal is to ensure that AI systems learn from human expertise without simply reproducing undesirable reasoning patterns.

Human-Centred Decision Support

Beyond improving AI models themselves, I explore intelligent decision-support systems that work alongside human experts. I am particularly interested in AI systems that can recognise when clinicians may be susceptible to heuristic-driven reasoning and provide timely, context-aware recommendations that encourage more reflective and evidence-informed decision-making.

Research Approach

My work combines statistical modelling, latent variable modelling, explainable AI, machine learning, and behavioural analysis to study the relationship between human cognition and artificial intelligence. Rather than viewing human decision-making and AI as separate challenges, I investigate how they interact throughout the AI lifecycle, from data generation and model development to deployment and human-AI collaboration.

Long-Term Vision

My long-term goal is to develop AI systems that are not only accurate and effective but also transparent, trustworthy, and aligned with the way people think and make decisions. I believe that the next generation of AI should do more than automate decisions, it should understand, complement, and support human reasoning, enabling safer and more informed decision-making in healthcare and other high-stakes domains.