Research Study

Personalised Audio Deepfake Detection

Choose a phase to take part. Each phase takes around 10–15 minutes and can be completed independently.

Phase 1 — Voice Enrolment
Record five short sentences so we can build a personalised voice profile for testing the detection system.
Approx. 10–15 minutes. Microphone required.
Start Phase 1
Phase 2 — Listening Test
Listen to 20 short audio clips and judge whether each one is a real or synthesised voice. Fully anonymous.
Approx. 10–15 minutes. Headphones recommended.
Start Phase 2