Extract from Cassandre Coyer’s article “Catch Deepfakes If You Can: Can E-Discovery Tools Keep Up With Gen AI?”
Blank eyes, alien-like hands or too many teeth—some of the AI-generated photos and videos have sometimes been pretty recognizable. But in other cases, voice clones and other types of hyper-realistic deepfakes have become almost impossible to discern with the human eye.
For e-discovery professionals standing on the first line of defense against inauthentic evidence making their way into courts, deepfakes are already raising concerns—especially since their tool stack may rapidly become outdated.
For now, e-discovery providers say there is no silver bullet solution on the market to detect generative AI-powered deepfakes. While it’s likely that vendors will soon upgrade existing tools or release new products to help detect fraudulent content, the cat and mouse game between deepfakes and e-discovery will mean professionals might have to operate solution-less for a while.
Still, providers note that it doesn’t mean that e-discovery experts will be defenseless against deepfakes. Instead, examining metadata and leveraging computer forensics could help fill in the gaps.
