Deepfakes & Content Provenance
Also written as Deepfake, Deepfake Detection, Content Provenance, C2PA, Content Credentials, AI Watermarking, SynthID
Deepfakes are AI-generated or AI-altered video, audio or images made to look real. The defences are detection models that spot fakes, provenance standards such as C2PA (Content Credentials) that record where a piece of media came from, and invisible watermarks such as Google's SynthID.
Think of it like
Forged banknotes, and the watermarks and security threads that let you check a real one.
Junior or senior?
Directly relevant to recruiting: real-time face and voice swaps are now used by fake candidates in remote video interviews. For engineers, senior means having built or tuned detection and knowing how quickly it goes stale as generators improve.
Ask them
“How did you check that your detection still worked against newer generation tools?”