The Rise of Celebrity Deepfakes: What You Need to Know

The Rise of Celebrity Deepfakes: What You Need to Know

Google searches for ‘celebrity deepfakes’ have skyrocketed over the past few years, along with more specific queries like ‘Ariana Grande Deepfakes’ or ‘Scarlett Johansson deepfakes”. Nowadays, the results page is mostly filled with articles about deepfakes, much like this one, but the truth is that most people are actually trying to find something much more sinister. 

In this article, we’re going to explore what celebrity deepfakes are, and the threats they pose both to the celebrities being targeted and to us common folk browsing the internet. 

What are celebrity deepfakes?

Simply put, celebrity deepfakes are fake AI videos of celebrities, usually created without their consent. They can portray your favorite celebrity doing or saying things they didn’t actually do, and sadly, they’re most often used to create scam videos or non-consensual pornography. 

What is a deepfake, and why does it matter? 

Deepfakes are an AI-based technique of synthesizing media that essentially allows people to superimpose someone’s face and features onto another person’s body, essentially creating very convincing fakes. The same principles apply to faking someone’s voice or body gestures. 

The reason why deepfakes are a bigger deal than traditional image and video manipulation, however, is that deepfakes can be a lot more convincing, and, more importantly, take very little effort to create. 

While manually faking a video using traditional video editing software would take weeks of work and require years of experience and practice, creating a deepfake takes as little as a couple of minutes and zero actual editing experience. 

This means that pretty much anyone can create a convincing celebrity deepfake video with the right app, and that misinformation, scams, and defamatory videos are more widespread than ever before. 

Ethical considerations and legal implications

Are celebrity AI videos ethical? 

We would be hard-pressed to argue that using someone’s likeness without their consent is in any way ethical. While some deepfakes, like memes which are created to entertain or satirize (and are clearly marked as AI-generated) are definitely less harmful than non-consensual pornography or scam videos, they are still more invasive than traditional parodies. 

Considering applicable laws

Since AI-generated content and deepfakes are still relatively new, legislation still differs wildly depending on the country. Most places have managed to get legislation in place to regulate the creation of deepfakes, but a lot of these laws tend to focus more on nudes and other types of non-consensual pornography and less on the use of deepfakes for defamation, scamming, and spreading misinformation. 

Famous celebrity deepfake videos 

Over the past few years, a number of notable celebrity deepfake videos have gone viral after they were spread on social media in an attempt to scam people or defame celebrities by associating them with controversial causes. 

Scarlett Johansson antisemitic deepfake

Earlier this year, a video featuring an AI-generated image of Hollywood actress Scarlett Johansson, among other Jewish celebrities including Jack Black, Mila Kunis, and Jake Gyllenhaal went viral on social media. In the fake video, Johansson was pictured wearing a shirt with the word ‘Kanye’ written on it, along with the Star of David and a hand giving the middle finger. 

The video was shortly published before Kanye returned to X (formerly known as Twitter) and started posting antisemitic messages. 

Johansson was quick to distance herself from the offensive deepfake, stating that “she has no tolerance for antisemitism or hate speech of any kind”. 

She also urged lawmakers to make the passing of legislation that regulates AI a top priority. In an interview with People magazine, the actress said she believes that “We must call out the misuse of A.I., no matter its messaging, or we risk losing a hold on reality.”

Johansson also sued the developer of an AI image editing app for using her voice in an ad without her permission back in 2023. Just a year later, she called out OpenAI for using a likeness of her voice in ChatGPT, causing them to stop using it. 

The Live demonstration of GPT4-o, featuring a voice eerily similar to Johansson’s.

Suggestive Facebook ad featuring deepfake Emma Watson

Back in 2023, a deepfake ad featuring Emma Watson ran on multiple Meta platforms, including Facebook, Messenger, and Instagram. The ad was part of a campaign featuring multiple unconsenting celebrities and promoted an app that promised users to swap anyone’s face onto any video of their choosing. 

Reporters at the time were quick to point out that “anyone” was not limited to celebrities, but could also include regular people, with high schoolers often falling victim to these types of malicious deepfakes. 

Brad Pitt and Keanu Reeves deepfake scams 

Recently, multiple women have fallen victim to a new subset of ‘pig butchering’ scams. The scammers used deepfake images and videos of celebrities like Brad Pitt or Keanu Reeves, as well as manipulative messages and tactics to build rapport with the victims before demanding money from them using various excuses. 

The Brad Pitt scammers, for example, claimed that the actor suffered from cancer and needed one million dollars to pay for his medical bills, saying that he couldn’t access his money due to his ongoing divorce from Angelina Jolie. 

These types of scams are nothing new. The perpetrators specifically target victims when they are at their most vulnerable, often preying on people who are going through difficult times. Before the rise of AI, impersonators would use random people’s stolen photos to pretend to be someone they weren’t, but the rise of AI has made it easier for them to impersonate celebrities and other public figures by generating ‘original’ images that can’t be found online, making it appear like the photos are genuine and taken specifically for the person on the other end of the conversation. 

Identifying and avoiding deepfake scams 

AI is clearly here to stay, and the number of scams just keeps going up. It’s pretty clear at this point that being able to reliably identify deepfakes has become a necessary skill if we want to survive in the modern digital landscape, and we’re here to give you some points that will help you do just that: 

Verify information before sharing it

Seen a video of a celebrity or public figure doing or saying something out of the ordinary? Look it up before sharing it! This is by far the quickest (and easiest) way to curb the spread of deepfakes. 

If the event is being reported on by multiple well-known publications, then it’s probably true. If it isn’t covered anywhere, then you probably encountered a deepfake. Do everyone a favor and report it if it’s misleading and isn’t clearly labeled as AI generated! 

PRO TIP: If you don’t want to waste time typing stuff into Google, you can just take a screenshot and then use circle-to-search (on Android) or the Google Lens app (on iOS) to look up the video or image you saw. If there’s any news coverage, it’s going to be on the first page of results. 

Approach everything with skepticism 

We know this is easier said than done, since scammers usually try their best to target us when we’re at our most vulnerable, but it’s still the most effective way to combat these types of attacks. 

Before you engage with someone pretending to be a celebrity, ask yourself if what’s happening makes sense. Why would your favorite movie star contact you of all people? Does it really make sense that someone with millions of dollars needs your money for their medical bills? 

Take a closer look at who posted it

Always check the accounts of the people who shared the video or contacted you. Is it verified, or is it just a random fan page? 
Looking up the celebrity’s name + your favorite social media network on Google will also usually reveal their official account, so you can see whether or not you’re talking to an impersonator.

Keep in mind that a lot of celebrities don’t actually have a social media presence. This can also be verified easily with a Google search, so if a public figure who clearly stated they don’t use social media is suddenly contacting you asking for help, you’ll know you’re being scammed.

Key indicators to identify deepfakes 

Deepfake technology is constantly evolving and getting better, but they’re still not fully indistinguishable from real footage. Here are some things to look out for when trying to determine whether or not a celebrity video you see online is a deepfake: 

1. Inconsistent lighting - look out for inconsistencies in the lighting and shadows that don’t make sense. 
2. Smoothed out facial features and blurry edges - When someone’s face is superimposed onto a video to create a deepfake, the results can often appear overly smooth and blurry, especially around the mouth or the edges of the face. 
3. Lips and audio aren’t sync - If the person’s lips aren’t moving according to what they’re saying, chances are that the audio was deepfaked. 
4. An odd number of fingers or teeth - While this has gotten better over the past year or so, AI still has a little difficulty figuring out how many fingers (or teeth) to give people. This nightmare-inducing detail is usually a surefire way to determine whether a piece of media was deepfaked. 
5. Slurred, stilted speech - AI voices can be pretty convincing, but they don’t always get it right. If your favorite celebrity sounds off in a video, this should tip you off that the audio might be faked. 

Celebrity deepfakes in digital and entertainment content 

Deepfakes can be dangerous, but doesn’t that mean that ALL celebrity deepfakes are malicious in nature. Over the last few years, there have been a number of pretty notable uses of legal, consensual deepfakes, usually in the entertainment industry. 

De-aging actors for movies and TV shows

Directors have used deepfake technology to de-age actors for a number of pretty big movies at this point, with varying degrees of success. One of the most talked-about examples in recent years has been the de-aging of Robert De Niro in The Irishman, but similar techniques have been used in Indiana Jones and Marvel’s Avengers films. 

Celebrities selling their likenesses to Meta

Back in 2023, Meta paid celebrities including Snoop Dogg, Kendall Jenner, and Charli D’Amelio millions of dollars to buy their likenesses and used their image to create AI chatbots that users could talk to on Instagram, Messenger, and Facebook. The bots did not attempt to impersonate the celebrities, but had their own personalities, and even their own accounts which were ran by the company, and they could carry conversations and answer questions.

The project was scrapped one year later, since the bot accounts were performing poorly and users were reportedly engaging with them less than they were engaging with the celebrities’ official accounts. 

However, the company doesn’t seem to be giving up on AI, with the company reportedly paying actors yet again to license their voices for an AI project that is soon to be announced, according to Bloomberg.

Raising awareness and charity campaigns 

Celebrity deepfakes have also been used in charity campaigns in order to help raise awareness about important issues. An official “Malaria Must Die” campaign featuring David Beckham deepfaked the footballer speaking in different languages in order to spread the message to different countries. 

Conclusion

In the end, it’s clear that celebrity deepfakes are neither a passing novelty, nor a fringe concern. They’re a clear sign of how AI is reshaping the digital landscape. What started as a technological breakthrough has quickly devolved into a torrent of scams, defamatory attacks, misinformation, and non-consensual content that targets both high-profile celebrities and everyday internet users. However, we’ve also seen how the technology can be harnessed for good, whether that’s creative storytelling, charitable campaigns, or cinematography. 

So, if AI and deepfake technology is here to stay,  then what can you do? First, you can arm yourself with a dose of healthy skepticism - you should always verify videos and images before you share, and rely on trusted sources whenever something feels off. Make sure to keep an eye out for odd facial details, mismatched audio, or jerky movements, and report any deepfakes that aren’t clearly labeled as AI-generated. Staying informed, demanding accountability, and using these technologies responsibly is the only way we can guide AI towards a future where innovation and integrity go hand in hand.