UX Collective

We believe designers are thinkers as much as they are makers. https://linktr.ee/uxc

Follow publication

Member-only story

AI-generated imagery: you might never be able to trust the internet again

Ari Joury, PhD
UX Collective
Published in
5 min readSep 28, 2022

Blue monster looking aghast at gray block which is printing a picture of a blue monster looking scary
AI-generated photos might have terrifying consequences for society. Image by author

EEver since the rise of deepfakes, citizens and governments have gotten increasingly concerned about fake photos and videos. Everyone seemed to be in danger: people with the right to vote might be target of political disinformation campaigns using deepfakes; people in court might be unable to distinguish a false piece of video evidence from a true one; people with jealous partners might find their faces appear in a porn clip — and struggle to explain that they never made that clip.

Some technologists dismissed the huge danger that deepfakes posed. Pictures have always been tampered with, they said, so what would make deepfakes more dangerous than Photoshop or the film industry in Hollywood?

The answer is their scale. Prior to the existence of deepfakes, editors had to spend hours to modify a single photo or video in a convincing way. Now, a computer can generate a deepfake in a matter of seconds.

This opened the floodgates for disinformation campaigns. Most recently, a deepfake got widespread attention in which Ukrainian president Zelenskyy told his soldiers to lay down his arms. It was quickly dismissed; but you can probably picture what might have happened if the Ukrainian government hadn’t reacted to it quicker.

Luckily for society, however, deepfakes are fairly easy to spot. The movement in these videos tends to be a little bit unnatural, and sometimes body parts don’t seem to fit together properly. Furthermore, deepfakes tend to be bad at showing people from a side angle or waving their hands in front of their faces.

All this might change with upcoming image generation technology.

Why deep learning image synthesis is so powerful

Deepfakes are useful to combine two or more pieces of media to form a new one. For example, if we wanted to generate a video of Barack Obama saying things he never said, we need a voice actor to make the soundtrack, and a short video of Obama saying something else.

Create an account to read the full story.

The author made this story available to Medium members only.
If you’re new to Medium, create a new account to read this story on us.

Or, continue in mobile web

Already have an account? Sign in

Written by Ari Joury, PhD

Founder of Wangari. Sustainable finance & ESG-financial modeling. Get all articles 3 days in advance: https://wangari.substack.com

Write a response

I remember mentioning to colleagues about a year ago that once people can generate images of themselves in whatever location, doing whatever they can imagine, a large portion of social media imagery will lose its power. People will divest from these…

--

Good read! You did some good points there.

--