Research Theme

Interactive & Applied AI

We codesign AI systems with domain experts, ensuring models serve real needs, respect constraints, and remain interpretable in the field.

Back to all themes View publications

Interactive AI in plain language

Interactive & applied AI means building machine-learning copilots that stand shoulder to shoulder with domain experts. We join doctors on ward rounds, sit in city-planning workshops, and shadow policy teams so we understand every constraint before we write a single line of code.

We focus on questions like:

How do we keep humans in the loop while the model adapts to new evidence?
What guarantees and explanations do stakeholders need before acting on a recommendation?
How do we ship solutions to edge devices or constrained environments without losing transparency?

Quick facts

Disciplines blended: HCI, ML engineering, responsible AI, design research.
Partners: Hospitals, mobility labs, climate initiatives, legal clinics, public institutions.
Deliverables: Trustworthy copilots, process maps, open-source toolkits, playbooks for decision-makers.

Field discovery

Participatory design, ethnography, and data audits reveal where AI can amplify human expertise without breaking critical workflows.

Human-centred optimisation

Preference learning, interactive labelling, and co-created dashboards keep performance, fairness, and usability aligned with stakeholder goals.

Responsible roll-out

Monitoring toolkits, governance reporting, and capability handovers ensure systems stay auditable and adaptable after deployment.

Key Publications

Conference and journal spotlights on human-in-the-loop deployments, preference learning, and responsible interfaces.

Browse the full publication library