The AI underbelly: Scandals, victims, and shocking celebrity deepfakes

Artificial intelligence, or commonly abbreviated as AI, had once only occupied a figment of human imagination, encapsulated in the works of science fiction and philosophical debate. Today, it is an increasingly prevalent aspect of our reality, especially through AI-powered digital photo and video enhancements that can be achieved in a matter of minutes. In fact, they have become so increasingly convincing, a new term has been coined to describe them: ‘deepfakes’, or AI generated videos using stolen identities that are almost indistinguishable from reality. And celebrities are increasingly becoming susceptible victims to deepfakes, leading to a dramatic rise in scandal and identity theft as a consequence of these AI-powered technologies.

When OpenAI’s ChatGPT first burst onto the helpbot and virtual assistant scene in November 2022, the world as we knew it was forever changed. Widely touted as the ‘next big thing’ in artificial intelligence, ChatGPT introduced users to a brand-new AI experience, almost immediately blowing competing chatbots clean out of the water.  

Empowered with the ability to interpret, refine, and respond to user inputs and provide human-like correspondence, ChatGPT as it stands today is the epitome of artificial intelligence. With its vast capabilities, authentic empathy, and seemingly boundless intelligence, ChatGPT quickly became the currency of the day, marking a pivotal shift in the use of AI in businesses, content creation, and the business of content creation. Those who mastered ChatGPT gained a valuable tool, enjoying assistance and improved efficiency across professional and personal spheres; those who did not were left fearing for their jobs and the future of humanity. 

By all accounts, ChatGPT recalls the wonders of Tony Stark’s AI assistants. First introduced in Marvel’s now-legendary MCU, J.A.R.V.I.S. and F.R.I.D.A.Y. were the creme de la creme of artificial intelligence, touted as the very best in a universe inhabited by Norse deities and Incredible Hulks alike. Yet, for all that Stark’s technology felt far beyond humanity’s reach, the day has now come that we can claim to be peers in terms of advancement, with AI having assimilated into its role in our everyday lives.  

The sobering reality, however, is far less inspiring. Technology has seen rapid advancements in recent years, far outpacing its earliest days in existence. This has sparked numerous debates amongst contemporaries, each voicing concern for the negative ramifications and potential threats posed by the unregulated expansion of AI-based systems.

The public is not alone in its mistrust of artificial intelligence. Last March, over 1,000 leaders and researchers in the technology field called for a pause on the development of the most advanced AI systems in an open letter, citing ‘profound risks to society and humanity’. 

This collective sentiment echoes that of key industry stakeholders, including tech billionaire Elon Musk, Apple co-founder Steve Wozniak, and Rachel Bronson, who is the president of the Bulletin of the Atomic Scientists, the body behind the Doomsday Clock. As per a letter from the nonprofit organisation Future of Life Institute, AI developers were “locked in an out-of-control race to develop and deploy even more powerful digital minds that no one — not even their creators — can understand, predict or reliably control.” 

In the realm of unbridled AI ambition, one question begs asking: How far is too far to go in pursuit of technological advancement? 

The ominous underbelly of artificial intelligence and deepfakes 

Despite its nonchalant assimilation into the modern world, the disquieting consequences of artificial intelligence have proven difficult to ignore. Of late, content across popular media and cautionary tales derived from real life incidents have provided a worrying glimpse into a world where AI is left to fester in the wrong hands. 

Most notably, the social ramifications of AI-generated deepfakes took centre stage in Netflix’s mind-bending dystopian anthology series Black Mirror. The season 6 premiere episode Joan Is Awful follows Joan Tait (portrayed by Annie Murphy) as she finds herself the sudden unwilling subject of an AI-generated show.  

As humans are wont to do, Joan is a woman of imperfection and sin who struggles with doing the right thing. However, things take a turn when she discovers a television show titled Joan Is Awful on the fictional streaming platform Streamberry (inspired by, and a parody of Netflix), in which her day-to-day life, deeds, and misdeeds are portrayed by actress Salma Hayek.

As a result of the show, Joan loses both her fiancé and her job, with her efforts to stop the show proving futile following the reveal that Hayek herself is powerless against the network’s usage of her likeness. 

The episode provides a vivid, horrifying, and thought-provoking caricature of generative AI its often-debilitating impact on society. From the eerie notion that our electronic devices are spying on us to the bone-chilling potential that unpoliced generative AI holds, Joan Is Awful provides unfettered commentary on the security and privacy threats we now face from within.

Unsettling as it is to watch the dire consequences of AI unfold on screen, its real-life parallels are equally alarming, if not more so. Of note is the recent dispute between actress Scarlett Johansson and OpenAI (of ChatGPT fame), as well as the latter’s development and use of a virtual AI personal assistant ‘Sky’ whose voice bears remarkable likeness to Johansson’s. While ‘Sky’ was first released to the public in September 2023, it was not until recently that it gained notoriety following a live demonstration by OpenAI. 

Observers’ comparisons between ‘Sky’ and Johansson hit close to home; Johansson had previously starred in Spike Jonze’s Her (2013), where she lent her voice to portraying a vastly intelligent AI assistant named Samantha. In a statement, Johansson revealed that she had been approached by OpenAI to voice the assistant but had declined to participate.

However, the actress has since expressed shock, anger, and disbelief over the release of ‘Sky’. Further fuelling the controversy, OpenAI CEO Sam Altman appeared to allude to Johansson’s portrayal of Samantha, with a tweet on May 13 that has since gone viral: ‘Her’. 

Johansson also addressed the tweet in her statement, saying, “Mr. Altman even insinuated that the similarity was intentional, tweeting a single word “her” – a reference to the film in which I voiced a chat system Samantha, who forms an intimate relationship with a human.” 

While OpenAI has since pulled the plug on ‘Sky’, Johansson rightly wants answers, with her legal counsel now demanding an explanation for how its development came to be. 

Outside the realm of celebrity and television, deepfakes have also made victims of streamers and online influencers. In January 2023, Twitch streamer Brandon “Atrioc” Ewing’s career came to a screeching halt when he accidentally revealed that he possessed explicit deepfake content of popular female Twitch streamers. Ewing’s undoing occurred while he was live-streaming; an errant click showed a tab on a pay-to-view site featuring deepfakes of his female streaming colleagues — several of whom he counted among his real-life friends. 

While Ewing has since issued a tearful apology, his career is unlikely to recover. Even more harrowing, however, is the resultant fallout for the female streamers involved, many of whom had not been aware that their deepfakes were being circulated in this manner.  

How to spot a deepfake: Top signs to look out for 

Once consigned to Photoshop experts, photo and video editing has become possible for the average layperson with the rise of AI-backed tools. Considering the ease with which one can erase imperfections in photographs, enhance videos, and even layer on AI-generated audio, and you’ll have the perfect recipe for disaster. From cases of identity theft to unscrupulous marketing involving AI-generated celebrity endorsements, deepfakes are fast taking over the world, claiming no small number of victims in its wake. 

Owing to the technological advancements of the day, many convincing deepfakes do exist on the internet. However, it is possible to discern reality from fantasy, and with enough practice, even the most gullible of internet users should be able to pick out something created using generative AI. According to industry-leading antivirus and security software company Norton, internet users can spot deepfakes by looking out for unnatural colouring, facial, body, and eye movements, and awkward body posturing and facial-feature positioning. Additional signs of generative AI include a distinct lack of emotions, unnaturally perfect hair and teeth, visual discrepancies like blurring and misalignments, and inconsistent audio. 

Celebrity deepfakes leave Taylor Swift, Oprah, and others victims of unregulated generative AI 

While it is not unheard of for laypeople to fall victim to deepfakes, the risk is amplified a thousandfold for celebrities who live their lives in the spotlight. From Taylor Swift and Maisie Williams to Gal Gadot and Emma Watson, numerous female celebrities had their likenesses utilised and manipulated for sexually explicit deepfakes, much of which remains unregulated on the internet. And while that warrants a deeper conversation into laws and regulations, it certainly cannot be done without raising more awareness on the subject. 

Here are some celebrities who have found themselves in the centre of deepfake scandals, from explicit viral smear campaigns on X (formerly Twitter) and questionable endorsements to personal data heists. None of it is true, and all of it is concerning. 

Taylor Swift 

You don’t become the most famous person in the world without suffering some fools, and in the case of Taylor Swift, many fools. The record-smashing singer-songwriter has proven beyond reasonable doubt that everything she touches turns to gold. This is perhaps most evident in the Eras tour; beyond clinching the honour of the world’s highest grossing tour, it has also generated billions in tourism revenue for hosting cities. And while the Chads and the Brads may complain about ‘Taylor Swift fatigue’, those in the business of marketing understand that the mere mention of her name is enough to rake in the views, clicks, and advertising dollars. 

In a world that hates to see an empowered woman succeed, it comes as no surprise that Taylor Swift has found herself unwillingly embroiled in multiple deepfake scandals. Perhaps the most prominent case of this was born out of her decade-long feud with controversial rapper Kanye West. Shockingly, a nude figurine of her was used, alongside the likeness of several other public figures, in West’s music video for Famous. Swift has famously referred to this depiction as ‘revenge porn’; yet, that would not be the only time the singer would suffer such an untoward attack on her person. 

Earlier this year, internet users of questionable morals created and circulated sexually explicit deepfakes of Swift at a football game. As expected, the images spread like wildfire on X (formerly Twitter), garnering over 45 million views and hundreds of thousands of likes, bookmarks, and reposts before they were ultimately removed 17 hours later. 

The attack prompted a statement from the Screen Actors Guild-American Federation of Television and Radio Artists which slammed the deepfakes, calling them “upsetting, harmful, and deeply concerning.” Echoing growing public unease about deepfakes, SAG-AFTRA continued, “The development and dissemination of fake images — especially those of a lewd nature — without someone’s consent must be made illegal.” 

Bobbi Althoff 

This past February, social media influencer and host of The Really Good Podcast Bobbi Althoff found herself the unfortunate subject of a deepfake scandal on X. The 26-year-old initially thought that she was trending on the social media platform because of a recent interview with rapper Wiz Khalifa. Unbeknownst to Althoff, she had become the latest name in celebrity victims claimed by explicit AI-generated deepfakes. 

While Althoff knew the images were fake, she later shared that the entire experience had opened her eyes to the damaging impact of generative AI, especially when placed in the wrong hands. Speaking to Yahoo Entertainment, Althoff shared that the deepfakes had fooled even her own team, who asked her, “We just need to ask: Is what we saw online real?” 

Unfortunately, X proved slow on the uptake to remove the circulating images. And while the immediate fallout has since died down, Althoff’s opinion is likely to transcend the years to come. “AI is scary,” She shared. “This world is scary. It’s getting scarier.” 

Ree Drummond 

One of the richest female chefs in the world, Pioneer Woman founder Ree Drummond has twice been the unfortunate victim of celebrity deepfakes. In 2022, she became the centre of controversy when AI-generated deepfakes emerged online, in which she promoted and endorsed the use of CBD gummies. While Drummond took to Facebook at the time to set the record straight, yet another deepfake scandal would erupt in December 2023 to sour her Christmas spirit. 

In the latest scandal, Drummond appeared in an AI-generated deepfake video to promote a fraudulent online campaign for Le Creuset. The video, which targeted social media users, offered freebies as a hook with which to steal the personal data and financial information of would-be participants. Drummond, Le Creuset, and the Federal Trade Commission have since come forward to denounce the video.

However, the Le Creuset deepfake provided a chilling glimpse into the sinister side of generative AI, extending beyond simply using a celebrity’s likeness for the endorsement of products. 

Oprah Winfrey, Nigella Lawson, and Piers Morgan 

In February, the deepfaked likeness of several celebrities including Oprah Winfrey, Nigella Lawson, and Piers Morgan were used in a series of online advertisements promoting a controversial self-help course by Wesley ‘Billion Dollar’ Virgin. Virgin, who calls himself a motivational coach, has 1.1 million followers on Instagram and claims to be worth USD 40 million. 

The self-help course, titled ‘Genie Script’, is touted as a ‘manifestation’ programme aimed at changing lives using allegedly ‘missing’ Hebrew Bible scripture, all for the low introductory price of USD 37.

In the promotional deepfakes, the AI-generated personifications of Winfrey, Lawson, and Morgan speak on how the programme has helped them to find success in their respective careers and personal lives.  

Responding to the deepfakes, a representative for the real Nigella Lawson came out to call the deepfakes ‘fraudulent’ and ‘of great concern’. Elsewhere, Piers Morgan told the BBC that the advert was “another example of a very worrying trend of public figures being misused by deepfake AI manipulators for financial gain”, adding thereafter, “The real victims will be members of the public who unwittingly buy these products believing the celebrity endorsements are genuine.” 

In yet another AI-generated video linked to the scandal, Oprah Winfrey appeared to praise the course, saying, “I want to give you a 20-word script… think of it like installing a new operating system into your mind that’s programmed to make you rich.” The video, which was discovered in the Meta ad library, showed genuine footage interspersed with AI-manipulated audio; however, it was easily proven to be a deepfake upon closer inspection. A spokesperson for the real Oprah has also denounced the video, expressing that the television personality had nothing to do with Virgin, nor his product.  

For his part, Virgin appears eager to discharge responsibility for the deepfakes. Speaking to the BBC, he said that the advertisements were the ‘work of affiliates’, and that he was ‘in the process of banning them all’.  

Rihanna, Lady Gaga, Dua Lipa, and Katy Perry, plus the Met Gala controversy 

With topics concerning classicism, international conflicts, and extreme poverty having reached boiling point in recent months, this year’s Met Gala was understandably inundated with drama and controversy. Add onto that the media firestorm surrounding social media influencer Haley Kalil (@haleybaylee), who sparked an avalanche of negative repercussions on social media in the wake of her Met Gala post. The post, which showcased her lavish Met Gala look, utilised a trending TikTok sound taken from Sofia Coppola’s Marie Antoinette (2006) in which the titular character blithely declared, in the face of starving peasants, ‘Let them eat cake’. 

While it cannot be proven that the real Marie Antoinette ever uttered those words, the tone-deaf TikTok post left a sour taste in netizens’ mouths, leading to Kalil’s swift social media crucifixion — but she would not be the only star to catch flack. In a movement now hailed as the ‘digital guillotine’, users on TikTok have begun blocking big-name celebrities, all in a bid to force them to acknowledge the ongoing horrors of the world. And while the jury is still out on whether Kalil’s influencing career will survive her grave faux pas, one important detail has emerged from the dregs of the evening: Society has become vulnerable in its gullibility. 

While neither Rihanna, Lady Gaga, Dua Lipa, nor Katy Perry attended the 2024 Met Gala, deepfake images of the celebrities soon showed up online, alluding to their presence at the event. Given the pristine and photoshopped nature of most red-carpet images, it made sense that the masses were fooled by the AI renderings. So convincing were these deepfakes, that Perry’s own mother reached out to offer congratulations on her spectacular Met Gala showing — only to be met by denial and amusement. After the incident, Perry recalled telling her mother, “Lol mom the AI got you too. BEWARE!” 

For all intents and purposes, the Met Gala deepfakes appear to be harmless. However, they certainly provide a bleak glimpse of the future, and for just reason. With the increasing spread and injudicious use of generative AI comes the unravelling of humanity’s social contract. Indeed, Perry herself has also addressed the potential risks of AI; earlier this year, she signed a petition alongside 200 other artists, urging technology platforms and digital music services to cease the use of AI to ‘infringe upon and devalue the rights of human artists.’ 

Morgan Freeman 

Last year, a very convincing deepfake of Morgan Freeman went viral online, sparking discussions on artificial intelligence and its place in the modern world. Posted by a Dutch deepfake channel DIEP NEP on YouTube, the video owes its concept and production to Bob de Jong, with voice acting by Boet Schouwink.  

Initially posted in 2021, the deepfake appeared to be education-driven, with Freeman’s AI avatar even announcing itself as a fake, starting off the video by saying, “I am not Morgan Freeman.” Nevertheless, it served as an interesting entry point into deeper discourse surrounding artificial intelligence and its use in creating believable human renditions.

(Main image: Taylor Swift/Instagram and Scarlett Johansson World/Instagram; featured image: Taylor Swift/Instagram)

Frequently Asked Questions (FAQs) 

– Who recently becomes the victim of deepfakes? 

Several celebrity personalities who have recently fallen victim to deepfakes include Taylor Swift, Bobbi Althoff, Nigella Lawson, Oprah Winfrey, and Piers Morgan. 

– Who are victims of deepfakes? 

Numerous celebrities have fallen prey to deepfakes, including Taylor Swift, Oprah Winfrey, Piers Morgan, Gal Gadot, Maisie Williams, Emma Watson, Rihanna, Katy Perry, and many others. 

– Which world famous person was featured in a deepfake image? 

Taylor Swift was recently the victim of explicit deepfake images which went viral on X (formerly Twitter). 

– What are celebrity deepfakes? 

Celebrity deepfakes take public figures and turn them into victims of identity theft. In essense, a celebrity deepfake is a digitally manipulated piece of media that is convincing, but ultimately fake. This media, which is often created using generative AI, can come in the form of videos, images, or audio. Deepfake creators also use the likeness of celebrities to add weight to their scams and endorsements; however, deepfakes also serve a more sinister purpose, targeting women specifically for the purpose of generating sexually explicit images to defame and degrade.  

This article first appeared here.