Human-AI Decision Making: Trust, Bias and Organisational Accountability

As AI tools become routine in finance, healthcare, logistics, recruitment, performance evaluation and public services, organisations face a compound challenge: getting humans and AI systems to work together in ways that are safe, effective and trustworthy, while managing the ways in which human and algorithmic biases interact, amplify each other and shape outcomes.

This project brings together two connected research strands: the dynamics of trust, boundary-setting and cognitive load in human-AI collaboration, and the interactions between human preconception and algorithmic bias in high-stakes organisational decisions. Together, they address one of the most urgent questions in contemporary management: how do we design and govern AI-augmented decision environments that are both effective and accountable?

Candidates may choose to focus on either of the following sub-directions, or to develop work that spans both:

Sub-direction 1: Trust, Cognitive Load and Human-AI Collaboration
*How trust in AI calibrates or miscalibrates in dynamic operational settings
*Patterns of cognitive load and mental-model alignment when humans and AI share decisions
*How AI explanations and confidence signals influence human inference and error
*How AI-generated omissions or framings reshape communication in team-based decisions
Sub-direction 2: Bias Interactions and Fair Decision Systems
*How human preconceptions shape the interpretation of algorithmic outputs (and vice versa)
*Override behaviours that intensify or compound flawed AI recommendations
*Design and evaluation of socio-technical interventions to reduce interactional bias
*Decision architectures and interfaces that help counteract classic human biases
Applicants should have a relevant degree in management, information systems, computer science, organisational behaviour or a related field. An interest in cognitive science, AI systems, human-machine collaboration or algorithmic fairness would be particularly useful.

Supervisors

Dr Chinomso Nwagboso

Demola Obembe