Protests and lawsuits have been increasing in recent months around the question of artificial intelligence (AI) training and copyright infringement claims. From authors, musicians, and actors to record company executives and Hollywood insiders, those involved in creative work are demanding that courts and governments address rampant violations of copyright law. AI companies claim fair use in training their AI models. Copyright holders vehemently disagree. Here’s the latest news.
Labels Push Back, Meta on Piracy Site, Protest Album
In February, the heads of Universal Music Group, Warner Music Group, and Sony Music Group mounted a Daily Mail campaign, where they opposed the UK Labor government’s controversial AI copyright exception proposal, which would loosen copyright laws for AI companies, and urged the UK government to protect the rights of creators.
Music Business Worldwide – Mandy Dalugdug – February 24, 2025
Also in February, it was brought to light that tech giant Meta, in training its AI models, used Library Genesis (LibGen), a shadow library that’s widely known as a major repository for pirated books. This information was first uncovered during a 2023 lawsuit that was filed by several authors that accused the tech company of copyright infringement. This revelation caused the US District Court for the Northern District of California to order the release of redacted court documents from the prior case, accusing the company of concealment.
ABP Life – Shayak Majumder – February 27, 2025
Over 1000 musicians, including Annie Lennox, Marillion, Simon Le Bon, Pet Shop Boys, and Kate Bush, released an album to protest the proposed British copyright law changes as they relate to AI. The album is titled Is This What We Want? It features studio silence and the sounds of static and human activities like footsteps and moving heavy objects, but no music or lyrics.
NPR – Chloe Veltman – February 27, 2025
Google Wants Fewer Copyright Rules, Hollywood and ASCAP Fight Back
March was a busy month in the fight between AI companies and creatives. A federal judge allowed part of a class-action lawsuit against Meta and OpenAI by comedienne Sarah Silverman and two authors to move forward. The lawsuit alleges that copyrighted books were used, without authorization, to train AI models. They also allege that Meta removed copyright information to hide the copyright infringement.
Later in the month, US District Judge Sidney Stein turned down OpenAI’s attempts to dismiss a copyright lawsuit filed by the New York Times that accused the company of mass copyright infringement of the paper’s content for AI system training without authorization or payment. The judge advanced the plaintiff’s core copyright infringement claims while narrowing the case’s scope.
Music Business Worldwide – Mandy Dalugdug – March 11, 2025
The Hollywood Reporter – Winston Cho – March 26, 2025
Also in March, Google and OpenAI published proposals asking the US government to allow them to use copyrighted content to train their AI models. These proposals follow the administration’s “AI Action Plan,” designed to enhance the country’s position in advanced AI technology while relaxing what it calls “burdensome requirements” that could impact innovation.
The Verge – Emma Roth – March 14, 2025
On March 17, Digital Trends reported that Google’s Gemini 2.0 AI model was able to remove watermarks from images that were copyright protected. The article states that while the company’s Gemini 2.0 Flash AI model won’t allow users to remove watermarks from images using its desktop or mobile version, they can successfully do it in Google’s AI Studio. This is a clear copyright violation that threatens the livelihoods of photographers and other artists.
Digital Trends / Yahoo! Tech – Nadeem Sarwar – March 17, 2025
Hundreds of Hollywood directors, actors, and other creatives are fighting against the relaxation of AI regulations. Over 420 entertainment industry workers signed an open letter to the federal government to uphold current copyright legislation regarding AI. Actress Natasha Lyonne led the group, which also includes Mark Ruffalo, Cate Blanchett, Ava DuVernay, Bette Midler, Ron Howard, Paul Simon, Aubrey Plaza, and many others. This letter is in response to Google and OpenAI’s recent request that copyright protections are lifted for artificial intelligence training.
Alden Global Capital, which owns dozens of newspapers, also slammed OpenAI and Google’s request for looser copyright laws, running editorials.
CBS News – Caitlin O’Kane – March 17, 2025
The Hollywood Reporter – Katie Kilkenny – March 18, 2025
Axios – Sara Fischer – March 17, 2025
The Verge reported that an Amazon Alexa Plus demonstration recently showed how the AI song generation platform Suno could be integrated into its application. Suno has admitted that its songs are pulled from copyrighted material but maintains that AI training usage is legal. The Amazon/Suno integration requires a Suno account before it can be linked to Alexa. With its integration demo, Amazon risks the ire of those companies that license music for its Alexa Echo speakers. They also undermine Suno’s fair-use legal argument.
The Verge – Elizabeth Lopatto – March 19, 2025
The American Society of Composers, Authors and Publishers (ASCAP) has joined the chorus of music industry professionals in opposing proposals to reduce copyright protections for AI development. The organization warns that if these copyright laws were changed, it would undermine the US’s creative economy.
Music Business Worldwide – Daniel Tencer – March 25, 2025
A California judge gave AI company Anthropic a small victory against music publishers by rejecting a preliminary bid that would have blocked it from using lyrics owned by UMG (Universal Music Group) and other music publishers. The judge said that the publishers failed to show how Anthropic caused them “irreparable harm.” The publishers, who claim that Anthropic uses lyrics without permission to train AI model Claude, remain confident that they will prevail in the long run.
Reuters – Blake Brittain – March 26, 2025
In the UK, prominent authors like Kate Mosse and Richard Osman signed an open letter to the UK government asking them to go after Meta for its illegal use of copyrighted material for its AI training. This follows the news that the tech company was using LibGen, which contains over 7.5 million books, without obtaining the authors’ permission. The letter, written by the Society of Authors (SoA) was published as a petition and has almost 5,000 signatures.
The Guardian – Ella Creamer – March 31, 2025
UK Architects vs AI Tech and OpenAI Infringement Update
In April, Top architects in the UK join musicians, actors, and other artists in calling for the government to protect intellectual property rights. The proposal, which they oppose, would allow AI firms like Midjourney and ChatGPT to create new content after training on copyrighted material without breaking copyright law. Creators would have to “opt out” to not be included in the AI training data.
Architects’ Journal – Gino Spocchia – April 2, 2025
A recent study suggests that OpenAI isn’t waiting for permission to flaunt copyright regulations. One of the three authors in the study is tech textbook giant Tim O’Reilly. The study shows that ChatGPT 4.o likely trained on 34 copyrighted O’Reilly Media books without the publisher’s permission.
The Register – Tobias Mann – April 3, 2025
UK creatives and lawmakers are fighting the government’s proposals for loosening AI copyright laws, requiring content creators to “opt out” of training usage or be subject to free use by tech companies, amounting to copyright infringement. However, AI companies want to go even further. OpenAI and Google have rejected the government’s proposals, saying that these new rules would bring implementation challenges and that the required transparency would be a burden.
Politico – Joseph Bambridge – April 3, 2025
A US judicial panel has decided to consolidate OpenAI’s lawsuit from news outlets and prominent authors. It has moved the California lawsuit, filed by comedian Sarah Silverman, author Ta-Nehisi Coates, and others to Manhattan federal court. It will be joined with cases from The New York Times and other news publications and authors such as Jonathan Franzen and John Grisham. OpenAI maintains its fair use argument in AI training, while the authors and news organizations accuse the company of copyright infringement.