ScarJo Condemns Deepfake of Her Protesting Kanye Ad As 'Misuse' of AI
PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.
Scarlett Johansson is speaking out after someone used her likeness without permission to create an AI-generated video of the actress criticizing Kanye West's antisemitic Super Bowl commercial.
The video shows an AI-generated Johansson and several other celebrities—Jerry Seinfeld, Mila Kunis, Jack Black, Drake—and tech founders Mark Zuckerberg, Sam Altman, Larry Page, and Sergey Brin wearing white T-shirts featuring a middle finger, a Star of David, and the word KANYE. "Enough is enough," the clip ends. "Join the fight against antisemitism."
West's Super Bowl commercial directed viewers to the website for his Yeezy clothing brand. But those who visited Yeezy.com instead saw only one item: a $20 T-shirt with a Swastika on it.
Johansson, who is Jewish, condemned West's commercial but also said there is no acceptable use of someone's likeness without their permission and urged the US government to pass legislation that protects its citizens against deepfakes. Johansson's agent provided her full statement to PCMag under the condition we would run it in its entirety:
"It has been brought to my attention by family members and friends, that an AI-generated video featuring my likeness, in response to an anti-Semitic view, has been circulating online and gaining traction. I am a Jewish woman who has no tolerance for antisemitism or hate speech of any kind. But I also firmly believe that the potential for hate speech multiplied by AI is a far greater threat than any one person who takes accountability for it. We must call out the misuse of AI, no matter its messaging, or we risk losing a hold on reality.
I have unfortunately been a very public victim of AI, but the truth is that the threat of AI affects each and every one of us.
There is a 1,000-foot wave coming regarding AI that several progressive countries, not including the United States, have responded to in a responsible manner. It is terrifying that the US government is paralyzed when it comes to passing legislation that protects all of its citizens against the imminent dangers of AI.
I urge the US government to make the passing of legislation limiting AI use a top priority; it is a bipartisan issue that enormously affects the immediate future of humanity at large.'
The deepfake video was originally posted by X user Ori Bejerano, who called out West for the Super Bowl ad as well as posts on X in which West declared "I am a Nazi," praised Hitler, and spread antisemitic conspiracy theories. It's still on X as of this writing; a Community Note about it being AI-generated was only added to a separate tweet sharing Bejerano's video.
This is the second time Johansson has spoken out about the misuse of AI. In May 2024, OpenAI released a demo of its AI-powered "Sky" voice assistant that sounded a lot like the actress. She accused the company of using her voice without permission, revealing that she turned down an offer from CEO Sam Altman to use her voice. OpenAI insists it found another voice actor, but it still disabled the Sky voice.
In both instances, there are sexual overtones to the use of Johansson's likeness. In the social media video about West's commercial, she appears in the first frame with tussled hair and a tight-fitting T-shirt. In the case of OpenAI, Altman wanted to use her voice to recreate the movie Her, in which the main character falls in love with Johansson's voice inside a computer.
Sexualized AI content has been a particular issue. In July, Meta failed to take down a naked AI photo of a female user in a timely fashion. High school students have also spoken out against the spread of explicit images of them at school, CNN reports. But AI-generated videos come in all shapes and sizes, especially as new tools make them free to produce. After the Kansas City Chiefs lost the Super Bowl, creators spun up humorous videos of quarterback Patrick Mahomes.
These videos are legal in the US, although and state lawmakers have introduced bills to go after unauthorized deepfakes. In September, California Governor Gavin Newsom that prevents AI deepfakes from appearing on social media ahead of an election. That prompted a , who argues it would make "parody illegal" on social media.
At this week's AI Action Summit in Paris, Vice President JD Vance endorsed an anti-regulation stance to AI development to promote the growth of the industry. 'We believe that excessive regulation of the AI sector could kill a transformative industry just as it's taking off,' Vance says. 'And we'll make every effort to encourage pro-AI policies.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time Magazine
an hour ago
- Time Magazine
How the Disney-Midjourney Lawsuit Could Reshape the Battle Over AI and Copyright
On Wednesday, the long-simmering dispute between Hollywood and the AI industry escalated dramatically when Disney and Universal sued Midjourney, one of the most prominent AI image generators, for copyright infringement. The two Hollywood heavyweight studios argue that Midjourney allows its users to 'blatantly incorporate and copy Disney's and Universal's famous characters,' such as Shrek and Spider-Man. 'Piracy is piracy, and the fact that it's done by an AI company does not make it any less infringing,' Horacio Gutierrez, Disney's chief legal officer, said in a general statement. The lawsuit challenges one of the AI industry's fundamental assumptions: that it should be allowed to train upon copyrighted materials under the principle of fair use. How the case gets resolved could have major implications for both AI and Hollywood going forward. 'I really think the only thing that can stop AI companies doing what they're doing is the law,' says Ed Newton-Rex, the CEO of nonprofit organization Fairly Trained, which provides certifications for AI models trained on licensed data. 'If these lawsuits are successful, that is what will hopefully stop AI companies from exploiting people's life's work.' A growing backlash against AI training norms AI companies train their models upon vast amounts of data scoured from across the web. While most of these companies have resisted admitting that they scrape copyrighted material, there are already dozens of AI copyright-related lawsuits in the U.S. alone alleging otherwise. Midjourney, which allows its millions of registered users to generate images from prompts, faces a class-action suit led by artists including Kelly McKernan, who found that users were inputting the artist's name as a keyword in Midjourney to spit out eerily similar artworks. 'These companies are profiting wildly off our unpaid labor,' they told TIME in 2023. For the last few years, Hollywood has refrained from entering the fray, while sending mixed messages about AI. During contract negotiations in 2023, AI was a major source of contention between unions like SAG-AFTRA and producers, who advanced a 'groundbreaking AI proposal' involving the use of 'digital replicas' to fill out the backgrounds of film scenes. But while some in Hollywood hope AI will make filmmaking more efficient and less expensive, many more have grown concerned about the AI industry's usage of copyrighted material. This concern has come to a head with the Disney-Universal lawsuit, which is the first major lawsuit brought by Hollywood studios against an AI company. The lawsuit seeks damages and an injunction that would immediately stop Midjourney's operations—and casts generative AI theft as a problem that 'threatens to upend the bedrock incentives of U.S. copyright law.' Midjourney did not immediately respond to a request for comment. 'We are bringing this action today to protect the hard work of all the artists whose work entertains and inspires us and the significant investment we make in our content,' said Kim Harris, executive vice president and general counsel of NBCU. Newton-Rex believes that this lawsuit is particularly significant because of the size, influence and resources of Disney and Universal. 'The more that these mainstays of the American economy weigh into this fight, the harder it is to ignore the simple truth here,' he says. In February, a Delaware judge dealt a blow to the AI industry's 'fair use' argument, ruling that a legal research firm was not allowed to copy the content of Thomson Reuters to build a competing AI-based legal platform. If the Disney-Universal lawsuit is similarly successful, that would have major implications for both AI and Hollywood, says Naeem Talukdar, the CEO of the AI video startup Moonvalley. Many AI companies might have to retrain their visual models from the ground up with licensed content. And Hollywood, if given legal clarity, might actually accelerate its usage of AI models built upon licensed content, like ones built by Natasha Lyonne's and Bryn Mooser's Asteria Film Co. 'Nobody wants to touch these models with a 10-foot pole, because there's a sense that you'll just get sued on the outputs later,' Talukdar says. 'I would expect that if this judgment falls a certain way, you'll see a lull, and then you'll have a new class of models emerge that pays the creators. And then you'll see this avalanche of studios that can now actually start using these models much more freely.' A governmental loophole? Unsurprisingly, AI companies are fighting back in court. They're also working on another path forward to retain their ability to train their models as they see fit: through governmental policy. In January, OpenAI sent a memo to the White House arguing their ability to train on copyrighted material should be ' preserved.' They then relaxed several rules around copyright in the name of 'creative freedom,' which triggered a flood of Studio Ghibli-style images on social media. In the U.K., the government announced plans to give AI companies access to any copyrighted work that rights holders hadn't explicitly opted out of, which drew a huge backlash from stars like Paul McCartney and Dua Lipa. Last week, the House of Lords rejected the legislation for a fourth time. Newton-Rex says that this dispute over AI and copyright will not be resolved any time soon. 'Billion-dollar AI companies have staked their entire businesses on the idea that they are allowed to take people's life's work and build on it to compete with them. I don't think they're easily going to give that up because of one lawsuit,' he says. Nevertheless, he says that the announcement of this lawsuit is 'really good for creators everywhere.'
TechCrunch
an hour ago
- TechCrunch
Apple is salvaging Image Playground with a boost from ChatGPT
When Apple first unveiled Image Playground last year, the company touted it as a quick and easy way to generate personalized, original images with the help of Apple Intelligence. Upon launch, the app was met with skepticism, as some users were disappointed by its limitations and low-quality results. For example, one Reddit user was surprised to find that Image Playground kept generating an image of a hand with six fingers when prompted to create an 'up close image of a hand.' In another instance, a user was unable to receive images of simple descriptions, such as 'old man' or 'flower.' Some people even questioned what the point of the app was, noting that it seemed juvenile when compared to popular AI image generators. (Of course, not everybody dislikes the app; some have described it as being a fun experience). Fast-forward to this year's WWDC, where the tech giant announced that it's going to boost Image Playground with a ChatGPT integration, which should allow for a better, more advanced AI image generation service. With the new ChatGPT integration, users will be able to access more styles that move beyond the emoji-like creations that Image Playground has become known for. Images can be created in the following new styles: Oil Painting, Watercolor, Vector, Anime, and Print. In addition, people will be able to select a new 'Any Style' option that allows them to describe exactly what they want. These new styles will be visible under a banner that reads 'ChatGPT style.' Image Credits:Apple In the past, users were limited to Animation, Illustration, Sketch, and Genmoji styles. Techcrunch event Save $200+ on your TechCrunch All Stage pass Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections. Save $200+ on your TechCrunch All Stage pass Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections. Boston, MA | REGISTER NOW Apple says that Image Playground will send a user's description to ChatGPT to generate an image. The tech giant notes that it won't share anything with ChatGPT without users' permission. By integrating ChatGPT into Image Playground, Apple is giving the app another shot at attracting users while positioning its AI image creation tool as a more well-rounded competitor to other similar free apps. Plus, it makes sense for Apple to leverage ChatGPT to boost Image Playground, especially since the company has already tapped OpenAI's tools in other ways. Last year, Apple announced that it was bringing ChatGPT to Siri and other first-party apps and capabilities across its operating systems. With Siri, the assistant leverages ChatGPT when it's unable to answer a question or its own or when a user specifically asks Siri to send an inquiry to ChatGPT. Apple has also integrated ChatGPT into system-wide tools like Writing Tools, allowing users to generate, rewrite, and summarize text in apps like Notes and Mail. The updated Image Playground app is expected to launch alongside the release of iOS 26 this fall.
Entrepreneur
an hour ago
- Entrepreneur
Disney, Universal Sue AI Startup Midjourney: 'Plagiarism'
The complaint alleges that Midjourney copied characters from the movie studios, including Darth Vader and Homer Simpson. Disney and Universal have brought the first major AI copyright lawsuit in Hollywood against AI image-generating startup Midjourney. In a 110-page complaint filed in U.S. District Court in Los Angeles on Wednesday, Disney and Universal accuse Midjourney of copying famous characters from their copyrighted works. The movie studios state that they have sent "cease and desist" letters to Midjourney's counsel to ask the startup to stop generating material featuring characters developed by the studios. Midjourney has allegedly disregarded their requests. "Midjourney, which has attracted millions of subscribers and made $300 million last year alone, is focused on its own bottom line," Disney and Universal allege. Related: A 74-Year-Old Needed a Lawyer, So He Used an AI Avatar in Court. It Didn't Go Well. Some of Disney's copyrighted characters include Darth Vader from "Star Wars," Elsa from "Frozen," and Homer Simpson from "The Simpsons," while characters from Universal include minions from "Despicable Me," Po from "Kung Fu Panda," and Hiccup from "How to Train Your Dragon." According to the lawsuit, only Disney and Universal are allowed to commercialize these characters and build a business around them. However, Midjourney has allegedly allowed its subscribers to generate images of characters like Darth Vader in violation of copyright laws. Disney and Universal included multiple examples in the complaint of AI-generated images from Midjourney featuring characters from "Cars," "Shrek," and other movies. Disney and Universal are asking for a jury trial, calling Midjourney's actions "textbook copyright infringement" and stating that the AI startup "threatens to upend the bedrock incentives of U.S. copyright law." "Midjourney is the quintessential copyright-free-rider and a bottomless pit of plagiarism," Disney and Universal allege. Related: New York Lawyer Uses ChatGPT to Create Legal Brief, Cites 6 'Bogus' Cases: 'The Court Is Presented With an Unprecedented Circumstance' Midjourney is a text-to-image AI generator that churns out images in seconds based on user prompts. It sells monthly subscriptions ranging from $10 per month for a basic plan to $120 per month for a mega subscription. The startup was founded in 2021 and has since generated $50 million in revenue in 2022 and $300 million in revenue in 2024. Midjourney notes on its website that it is "a small self-funded team" with "11 full-time staff." While Disney and Universal's lawsuit against Midjourney represents the first major Hollywood lawsuit against an AI startup, another groundbreaking AI case was filed last week. Reddit became the first major tech company to sue an AI startup, alleging in the complaint that the $61.5 billion startup Anthropic used the site for training data without permission. AI copyright cases can get expensive, too. Getty Images CEO Craig Peters said last month that Getty has spent millions of dollars in a years-long legal fight with AI image generator Stability AI. Getty alleged that Stability AI illegally scraped more than 12 million copyright-protected media from its site to train its AI image generator. Getty launched the suit in January 2023; the case is set for an initial trial on June 9.
