What those Taylor Swift deepfakes reveal about the future of AI

The Medium Newsletter
The Medium Blog
Published in
Sent as a

Newsletter

3 min readFeb 5, 2024

🤖 Morning! AI did not write this email, we swear.

What should we call the photorealistic images AI makes? Photos? And if they’re not photos, what are they? (“Images” is vague. The English language can’t keep up with whatever AI is doing… if you have suggestions, let us know.)

An AI researcher and photographer on Medium, Thomas Smith, argues that, no, these are not photos. A photo, he writes, is “a moment of light, captured, fixed, and frozen forever.” (The word “photography” literally means “drawing with light.” That’s not what AI is doing.)

Maybe you heard about those explicit AI-generated “photos” of Taylor Swift that were viewed more than 47 million times last month — and prompted X to briefly ban searches for “Taylor Swift.” The creator of those deepfakes? Probably a stable diffusion model similar to Midjourney 6, which came out in December. While the artist known as Midjourney 5 kept getting humans wrong (remember all those messed up hands?), Midjourney 6 makes fake people that don’t actually look fake. On Reddit, a prompt engineer did a side-by-side comparison:

Midjourney 5.2 v. Midjourney 6 (source). Notice those minute details, like freckles? That’s how our brains know it’s “human” (even though it’s not).

It seems innocuous (and a little eerie), but as Smith predicts:

[Midjourney 6] presents a huge problem for people who make a living creating simple, aspirational photos of people. Hiring a model […] purchasing wardrobe, installing appropriate lighting, and buying pro camera gear is expensive, and takes weeks. Conjuring up a fake [person] with AI costs nothing at all, and takes about 15 seconds.

Goodbye to the stock photo industry perhaps, but also hello to the fake celebrity photo industry. Eesh.

The thing about any new tool is that it can be used to enhance and amplify humanity — or to diminish it. 63% of workers who use generative AI already say it makes them more productive. How do you use AI, and how do you think we’ll use AI-generated imagery in the future? Leave a response on this story to let us know.

And if you’re a machine learning engineer or AI expert, what do you know that the rest of us don’t? Write about it on Medium.

What else we’re reading

  • One writer explains why André 3000, the 48-year-old rapper and singer who pivoted to flute, is the epitome of creative freedom: “So far, I’ve spent the better part of my 40 years being inspired by people putting their mind to being the best in their chosen field,” writes joahspearman, who attended a livestreamed André 3000 concert last week, “But what I hadn’t yet seen is what exactly that earns a person. I had yet to see exactly what true freedom could look like — beyond the notion of ‘f-you money.’”
  • A few years ago, engineer and AI researcher François Chollet published a widely-read list of 13 notes to himself about building good software. You can apply most of them to any field, like this one: “Ask not what your pull request can do for your next promotion, ask what your pull request can do for your users and your community. Avoid ‘conspicuous contribution’ at all cost. Let no feature be added if it isn’t clearly helping with the purpose of your product.”

Your daily dose of practical wisdom

There are few things more liberating in life than giving up the need to be liked by everyone.

Written by Harris Sockel
Edited and produced by
Scott Lamb & Carly Rose Gillis

Every weekday, The Daily Edition offers useful, human perspectives to help you become a better reader, writer, and thinker.

Want to receive this as an email? Sign up here.

Want to browse the best of Medium? Explore staff picks.

Questions, feedback, or story suggestions? Email us: tips@medium.com

--

--