Understanding Deepfakes and the Taylor Swift Images: QuickTake

The world is awash in deepfakes — video, audio or images that make people appear to do or say things they didn’t, or be somewhere they weren’t. Many are devised to give credibility to falsehoods and damage the reputations of politicians and other people in the public eye. But most deepfakes are explicit videos and pictures concocted by mapping the face of a celebrity onto the body of someone else. That’s what happened in late January, when fake explicit images of pop star Taylor Swift cascaded across social media. Now that artificial intelligence allows almost anyone to conjure up lifelike images and sound with a few taps on a keyboard, it’s getting harder to tell if what you see and hear online is real.

1. What happened to Taylor Swift?

The phony images of Swift were widely shared on social media sites, drawing the ire of her legions of fans. One image shared on X, the site formerly known as Twitter, was viewed 47 million times before the account was suspended, the New York Times reported. X said it was working to remove “all identified images” and would take “appropriate action” against those who posted them. Swift was also among the celebrities whose voices and images were manipulated into appearing to endorse commercial products — a popular brand of cookware, in Swift’s case.

2. Where else have deepfakes been in the news?

Earlier in January, Xochitl Gomez, a 17-year-old actress in the Marvel series, spoke out about finding sexually explicit deepfakes with her face on social media and not succeeding in getting the material taken down, NBC News reported. Deepfakes are also popping up in the 2024 US presidential election. New Hampshire residents received a robocall before the state’s presidential primary that sounded like President Joe Biden urging them to stay home and “save your vote for the November election.” The voice even uttered one of Biden’s signature phrases: “What a bunch of malarkey.”

3. How are deepfake videos made?

They are often crafted using an AI algorithm that’s trained to recognize patterns in real video recordings of a particular person, a process known as deep learning. It’s then possible to swap an element of one video, such as the person’s face, into another piece of content without it looking like a crude montage. The manipulations are most misleading when used with voice-cloning technology, which breaks down an audio clip of someone speaking into half-syllable chunks that can be reassembled into new words that appear to be spoken by the person in the original recording.

4. How did deepfake technology take off?

The technology was initially the domain of academics and researchers. However, Motherboard, a Vice publication, reported in 2017 that a Reddit user called “deepfakes” had devised an algorithm for making fake videos using open-source code. Reddit banned the user, but the practice spread. Initially, deepfakes required video that already existed and a real vocal performance, along with savvy editing skills. Today’s generative AI systems allow users to produce convincing images and video from simple written prompts. Ask a computer to create a video putting words into someone’s mouth and it will appear. The digital forgeries have become harder to spot as AI companies apply the new tools to the vast body of material available on the web, from YouTube to stock image and video libraries.

A fake image of Pope Francis generated by Midjourney AI, left, and a real photo of Pope Francis.

Source: Midjourney; AFP/Getty Images

5. What are some other examples of deepfakes?

Chinese trolls circulated manipulated images of the August wildfires on the Hawaiian island of Maui to support an assertion that they were caused by a secret “weather weapon” being tested by the US. In May 2023, US stocks dipped briefly after an image spread online appearing to show the Pentagon on fire. Experts said the fake picture had the hallmarks of being generated by AI. That February, a manufactured audio clip emerged with what sounded like Nigerian presidential candidate Atiku Abubakar plotting to rig that month’s vote. In 2021, a minute-long video published on social media appeared to show Ukrainian President Volodymyr Zelenskiy telling his soldiers to lay down their arms and surrender to Russia.

6. What’s the danger here?

The fear is that deepfakes will eventually become so convincing that it will be impossible to distinguish what’s real from what’s fabricated. Imagine fraudsters manipulating stock prices by producing forged videos of chief executives issuing corporate updates, or falsified videos of soldiers committing war crimes. Politicians, business leaders and celebrities are especially at risk, given how many recordings of them are available. The technology makes so-called revenge porn possible even if no actual naked photo or video exists, with women typically targeted. Once a video goes viral on the internet, it’s almost impossible to contain. An additional concern is that spreading awareness about deepfakes will make it easier for people who truly are caught on tape doing or saying objectionable or illegal things to claim that the evidence against them is bogus. Some people are already using a deepfake defense in court.

7. Is anything being done about it?

The kind of machine learning that produces deepfakes can’t easily be reversed to detect them. But a handful of startups such as Netherlands-based Sensity AI and Estonia-based Sentinel are developing detection technology, as are many big US tech companies. Intel Corp. launched a FakeCatcher product in November 2022, which it says can detect faked video with 96% accuracy by observing the subtle color changes on the subject’s skin caused by blood flow. Companies including Microsoft Corp. have pledged to embed digital watermarks in images created using their AI tools in order to distinguish them as fake. US state legislatures have moved faster than Congress has to tackle the immediate harms of AI. Several states have enacted laws that regulate deepfakes, mostly in the context of pornography and elections. A proposed European Union AI Act would require platforms to label deepfakes as such.

The Reference Shelf

• Related QuickTakes on generative AI and AI regulation.
• How Google and Microsoft are supercharging AI deepfake pornography.
• Bloomberg Law says US regulators are wrestling with how to stop deepfakes affecting the 2024 presidential election.
• A Bloomberg video about Lyrebird, the AI company that puts words in your mouth.
• Research from University College London suggested humans were unable to detect more than a quarter of deepfake audio recordings.

--With assistance from Jillian Deutsch, Diana Li, Isaiah Poritz, Laurence Arnold, John O’Neil and David S. Joachim.

To contact the reporter on this story:
Nate Lanxon in London at lanxon@icloud.com

To contact the editors responsible for this story:
Lisa Beyer at lbeyer3@bloomberg.net

Thomas Pfeiffer

© 2024 Bloomberg L.P. All rights reserved. Used with permission.