Help train a personalised deepfake detector

This study is testing a system that detects computer-generated (deepfake) audio. You'll record yourself reading five short sentences so the system can learn the sound of your voice. Your recordings are then used to test whether it can tell your real voice apart from synthesised versions of it.

Your data is stored under an anonymous identifier — never linked to your name. Recordings and generated audio are deleted within 14 days of the study ending.