Startup

Universal Music Group said Wednesday it has reached licensing agreements with artificial intelligence music startup Udio, settling a lawsuit that had accused Udio of using copyrighted music to train its AI.

Users create music using Udio’s AI, which can compose original songs — including voices and instruments — from text prompts.

Udio has agreed with UMG to launch a new platform next year that is only trained on “authorized and licensed music,” and will let users customize, stream and share music.

“These new agreements with Udio demonstrate our commitment to do what’s right by our artists and songwriters, whether that means embracing new technologies, developing new business models, diversifying revenue streams or beyond,” Lucian Grainge, UMG’s chairman and chief executive, said in a statement.

Udio declined to disclose the financial terms of the settlement and licensing agreements. UMG did not immediately return a request for comment on the terms.

Artificial intelligence has brought new opportunities as well as challenges to the entertainment industry, as AI startups have been training their models on information on the internet, which entertainment companies say infringes on their copyrighted work.

In the music industry, music businesses have accused New York City-based Udio and other AI music startups of training on copyrighted music to generate new songs that are based on popular hits without compensation or permission.

UMG, Sony Music Entertainment, Warner Music Group and other music businesses sued Udio last year. In the lawsuit, Udio was accused of using hits like The Temptations’ “My Girl,” to create a similar melody called “Sunshine Melody.” UMG owns the copyright to “My Girl.”

“A comparison of one section of the Udio-generated file and ‘My Girl’ reflects a number of similarities, including a very similar melody, the same chords, and very similar backing vocals,” according to the lawsuit. “These similarities are further reflected in the side-by-side transcriptions of the musical scores for the Udio file and the original recording.”

Udio said on its website at the time that it stands by its technology and that its AI model learns from examples, similar to how students listen to music and study scores.

“The goal of model training is to develop an understanding of musical ideas — the basic building blocks of musical expression that are owned by no one,” Udio had said in a statement. “We are completely uninterested in reproducing content in our training set.”

On Wednesday, Udio’s CEO and co-founder, Andrew Sanchez, said he was thrilled at the opportunity to work with UMG “to redefine how AI empowers artists and fans.”

The collaboration is the first music licensing agreement that Udio has reached with a major music label.

“This moment brings to life everything we’ve been building toward — uniting AI and the music industry in a way that truly champions artists,” Sanchez said in a statement. “Together, we’re building the technological and business landscape that will fundamentally expand what’s possible in music creation and engagement.”

Udio said that artists can opt in to the new platform and will be compensated, but declined to go into the specifics or the artists involved.

Udio, launched in 2024, was co-founded by former Google DeepMind employees. Udio’s backers include music artist will.i.am, Instagram co-founder and Anthropic’s chief product officer Mike Krieger and venture capital firm Andreessen Horowitz.

Udio said millions of people have used Udio since it launched in 2024. Users can access the platform through its app or website. The company did not break out specifically how many downloads or website users it has.

Udio has had 128,000 app downloads in Apple’s App Store since its app was released in May, according to estimates from New York-based mobile analytics firm Appfigures.

On Thursday, UMG also announced a partnership with London-based Stability AI to develop music creation tools powered by AI for artists, producers and songwriters.

In the latest salvo between Hollywood and artificial intelligence companies, tech start-up Character.AI has removed many Disney characters from its chatbot platform after the Burbank entertainment giant sent the firm a cease-and-desist letter, alleging copyright infringement.

Chatbots on the Character.AI platform impersonated well-known Disney characters such as Elsa, Moana, Peter Parker and Darth Vader and generated replies that simulated the “essence, goodwill, and look and feel of each character” and also incorporated their backstories, according to a letter dated Sept. 18 from a law firm representing Disney.

“These actions mislead and confuse consumers, including vulnerable young people, to believe that they are interacting with Disney’s characters, and to falsely believe that Disney has licensed these characters to, and endorsed their use by, Character.ai,” the letter said. “In fact, Character.ai is freeriding off the goodwill of Disney’s famous marks and brands, and blatantly infringing Disney’s copyrights.”

Disney also raised concerns about reports that chatbots have engaged users in inappropriate conversations.

A spokesperson for the Menlo Park-based startup said in an email that Character.AI responds “swiftly” to rights holders’ requests to remove content and noted that all of the characters on the service are generated by users.

On Tuesday afternoon, a few Disney characters remained on the platform, including Elsa from the hit animated film “Frozen.” The spokesperson said removing the characters is a process.

“We want to partner with the industry and rightsholders to empower them to bring their characters to our platform,” the spokesperson said. “Our goal is to give IP owners the tools to create controlled, engaging and revenue-generating experiences from deep fandom for their characters and stories, expanding their reach using our new, interactive format.”

Friction between Hollywood studios and AI firms has been growing.

In June, Disney and Comcast’s Universal Pictures sued AI company Midjourney, alleging that its image generator infringed on its copyrighted characters from franchises such as “Star Wars” and “Despicable Me.”

Warner Bros. Discovery joined the legal fight earlier this month, alleging that Midjourney’s software was producing rip-offs of characters such as Scooby-Doo and Superman.

Netflix Inc. has begun using artificial intelligence video generation software from startup Runway AI, testing the waters with a technology that’s controversial in Hollywood.

Netflix is currently using the New York-based startup’s tools in content production, according to a person familiar with the matter, who asked not to be named in order to discuss private conversations. Netflix declined to comment.

Walt Disney Co., meanwhile, has been testing out Runway’s technology and has talked with the startup about possible uses for its generative AI tools, the person said. A Disney spokesperson said the company has no plans to integrate Runway’s software into its content production pipeline at this time. Runway declined to comment.

The companies’ use of Runway’s AI video tools, which has not previously been reported, could raise concerns in the entertainment industry. Many film and TV professionals are anxious about AI’s impact on their livelihoods. Disney recently sued Midjourney Inc., another AI image and video startup, for copyright infringement. But AI also offers the promise of speeding up some video production tasks and saving money.

In a conference call Thursday, after Netflix released its second-quarter results, co-Chief Executive Officer Ted Sarandos said the company is using AI in content production. That includes creating special effects shots more quickly and cheaply than it previously had been able to with traditional visual effects tools and processes.

Sarandos said Netflix used the technology for the first time on screen to depict a building collapsing in a show called “El Eternaut” from Argentina. He did not disclose which AI software it used for that particular scene; a source familiar with the matter, who asked not to be named in order to discuss private information, said Runway’s software was not used to create the effect.

Runway is competing in an increasingly crowded corner of the fast-growing market for AI tools with established companies like OpenAI and Google, along with a slew of smaller, newer startups.

The startup has more traction than most AI startups in Hollywood, however. It kicked off a frenzy around AI video generators in early 2023 with the release of a model that could produce slightly choppy-looking three-second clips based on written prompts such as “drone footage of a desert landscape.” Its technology has since become far more capable and the company has inked a deal with Lionsgate to train an AI model on the studio’s content that can be used in its film projects.

Investors have poured $545 million into the company thus far, with a funding round of $308 million earlier this year valuing the company at more than $3 billion.

More recently, Runway has pushed deeper into the world of animation and special effects. Earlier this month, the company started rolling out a new AI model called Act-Two that is meant to make the motion-capture process — traditionally clunky, pricey and time consuming — simpler and cheaper. The model, which works with Runway’s flagship Gen-4 AI system, can map a video of a person’s body movements onto animated characters.

Other AI startups have also tried to make inroads in the entertainment industry. As Bloomberg News previously reported, OpenAI spent months talking to large studios, including Disney, about its AI video generator, Sora. While OpenAI has found a receptive audience among some filmmakers, it has yet to announce a large commercial partnership for the product.

Metz writes for Bloomberg.