NeRF: An Eventual Successor for Deepfakes?

This story was initially published on Metaphysic.ai by Martin Anderson.

In February of 2022, a study from Lancaster University in the UK found not only that most people cannot distinguish deepfake faces from real faces, but also that we tend to trust synthesized faces more.

While this is excellent media fodder, and while there’s little doubt that AI-generated humans have increased in quality since the advent of deepfakes in 2017, it’s worth considering the extent to which our discriminatory powers and standards for authenticity are likely to rise in line with improvements in deepfake generation; that there’s still a long way to go before machine learning can reproduce humans with complete fidelity; and that the technologies that are currently dazzling us are not necessarily the ones that will achieve the next evolutionary leap in this sphere.

Most of the attention-garnering headlines regarding deepfakes are referring to the use of two open source packages that became popular after deepfakes entered the public arena in 2017: DeepFaceLab (DFL) and FaceSwap…

To read the full story please click here.