CYBER & DIGITAL TRUST
Deepfake & media detection
To recognize the synthetic, start with the human.
Authenticity systems need reference examples of genuine media and carefully labeled manipulations.
The human contribution
Consented recordings of people speaking and acting across capture conditions, paired with separately labeled synthetic or edited examples.
The structured dataset
Audio and video samples, source labels, edit histories and separated evaluation sets.
Where it helps
Train and evaluate systems that flag manipulated media, voice impersonation and fraudulent content.
Collection scope, availability, participant permissions and permitted use are defined for each engagement.
Discuss a data program ↗