I don't think so, I tested and neither detector nor recognizer were fooled. Here are the results with FaceNet, Inception, MTCNN face detector and MobileNet SSD detector. Facenet also retrieved similar looking faces from LFW.

[1] https://www.dropbox.com/s/weasqs8pxn9q7qv/screencapture-loca...

[2] https://www.dropbox.com/s/i2nii4biy8aioyo/screencapture-loca...

[3] https://www.dropbox.com/s/lmq8ldqujyk3pa9/screencapture-loca...

I guessed as much but their end product is presumably a face image that is obfuscated enough to confuse a recognition algorithm, like AWS's Facial Recogniton API: https://aws.amazon.com/rekognition/details/

Note: The AWS endpoint for face detection still recognizes the objects as faces, but that's to be expected. It'd be pretty damn amazing for an obfuscation algorithm to fool object detection.

Would love to see how what D-ID's algorithm does to Obama's presidential portrait, and then to test it against a trained AWS Rekognition model on Obama's Flickr account.

edit: FWIW, the results of AWS's face detection on the D-ID headshots: http://imgur.com/a/cPsQK

Again, D-ID isn't claiming they can evade face detection, but face recognition. Being able to do the former (while maintaining integrity of the image) would be a pretty mindblowing feat :)