Investigative interviewing, studied across law, psychology, neuroscience and technology

In this MRQ we will develop and deploy an instrument which measures compliance with the Méndez Principles within individual jurisdictions. We will gather information from around the globe to create an open access interactive online map of Méndez compliance. In doing so we will examine systems and people, statute law, law in action, and the potential need for local adaptations.

A second study will involve a closer look at law and practice in certain jurisdictions, examining police methods and procedural safeguards. This will be an empirical study, engaging with serving and retired police officers, public prosecutors (where relevant), lawyers, and academic experts to gain insights into the justifications for the interview and interrogation techniques in use and the extent of any existing safeguards.

Later in the project life-cycle we will undertake Action Studies in a smaller number of jurisdictions, observing decision-making, planning and interviewing strategies and techniques in the context of simulated, but naturalistic, mock investigations. We will review these Studies in conjunction with police supervisors and legal experts across jurisdictions, providing an opportunity for cross-jurisdictional sharing of expertise, experience, and reflection on the application of the Méndez Principles.

The final aspect of this MRQ is engagement with policy-makers internationally, to explore existing knowledge and application of the Méndez Principles and to chart a way forward in terms of enhanced compliance, where possible.

MRQ 1 aims to provide an understanding of the conditions required for Méndez compliance and the basis for a ‘Roadmap to Méndez’ to guide states, systems, and people away from coercive and abusive practices. The studies within MRQ1 will provide clarity on the state of police interviewing, and suspect safeguards, worldwide and increase a sense of momentum for the implementation of the Méndez Principles.

MRQ2 examines the underlying mechanisms that drive individuals and systems to either adopt or resist coercive and abusive interviewing practices. We will investigate the neural, psychological, and social processes that shape decision-making in high-stakes investigative contexts. Through a series of experimental studies, participants will engage with realistic scenarios to explore factors such as silence under pressure, diffusion of responsibility, the influence of social norms and authority, and trust in coerced information.

Using behavioural experiments, group-based tasks, and advanced neuroimaging methods (including fNIRS hyperscanning), we will capture how individuals think, decide, and interact in these contexts. By integrating behavioural, neural, and narrative data, we will develop a comprehensive understanding of how coercive practices are justified, normalised, or resisted in real-world settings.

MRQ2 aims to uncover the core mechanisms that sustain coercive practices across individuals and systems, providing a critical foundation for designing interventions that promote ethical, human rights-compliant decision-making.

MRQ3 explores the impact of coercion on people who are subjected to it. Two studies will approach this issue from different perspectives.

This first will look at the short-term impact of coercive and abusive interview practices. We will conduct experiments to examine the impact of seeing someone else experiencing stress: does this linger in the mind of witnesses and affect or direct the course of their cognition?

Secondly, we will explore the longer-term impact of coercive and abusive interview practices on those subjected to them. We will carry out semi-structured, qualitative interviews with 20 people across various jurisdictions who have been the subjects of coercive interviewing. Researching the impact on their psychological wellbeing will give voice to the lived experience of interviewees in investigative settings and allow us to develop a typology of individualised responses to coercive and abusive interview practices.

New knowledge on the long-term lived experience and short-term consequences of coercive and abusive interviewing techniques will be discovered under MRQ3.

MRQ4 focuses on developing and testing innovative, evidence-based strategies to promote human rights-compliant investigative interviewing. Central to this effort is the creation of a novel “Méndez LLM” — a large language model specifically designed to support ethical, non-coercive interviewing in line with the Méndez Principles.

We will train and validate this system relying on two key computational methods to identify effective interviewing strategies and refine it through large-scale interactions with both models and human participants. We incorporate reinforcement learning, generative language models and adaptive model optimisation, addressing a fundamental scientific challenge: how can highly flexible generative models be aligned with explicit ethical constraints in complex conversational settings? Answering this question will enable continuous learning and optimisation of interviewing approaches strictly within ethical constraints. Subsequent behavioural experiments will evaluate how such tools can be integrated into real-world investigative practice, exploring hybrid intelligence approaches where human and computational intelligence are meaningfully combined.

MRQ4 aims to deliver practical tools and actionable strategies that can transform investigative interviewing worldwide, establishing a new standard for ethical, evidence-based practice and accelerating the global adoption of the Méndez Principles.