AI-generated Fake Faces Seem More Trustworthy Than Real Faces

AI-powered deep fake technology is improving rapidly. A recent study in the Proceedings of the National Academy of Sciences found that existing software generates artificial faces that are so convincing that viewers can't distinguish them from real people. Moreover, these artificial faces tend to be regarded as appearing to be more trustworthy than actual human faces.

This finding may have enormous interest to those who use the technology for commercial or political purposes, and which of course may have enormous impact on all of society.

The UK/US team of researchers who reported this startling finding included a set of recommendations to respond to this immensely powerful tool. They wrote:

"We, therefore, encourage those developing these technologies to consider whether the associated risks are greater than their benefits. If so, then we discourage the development of technology simply because it is possible."

Link to full-text versions of the study in HTML and in PDF:

Sophie J. Nightingale, and Hany Farid. (2022). AI-synthesized faces are indistinguishable from real faces and more trustworthy. Proceedings of the National Academy of Sciences Feb 2022, 119 (8) e2120481119; DOI: 10.1073/pnas.2120481119