AI: Google and Universal negotiate agreement on ‘deepfakes’ – 09/08/2023 – Tech

Google and Universal Music are in talks to license melodies and artist voices for AI-generated songs as the music industry tries to monetize one of its biggest threats.

The talks, confirmed by four people familiar with the matter, are aimed at forging a partnership in an industry grappling with the implications of new AI technology.

The rise of generative AI has spawned a wave of “deepfake” songs that can convincingly imitate the voices, lyrics or sounds of established artists, often without their consent.

Frank Sinatra’s voice was used in a version of the hip-hop song “Gangsta’s Paradise”, while Johnny Cash’s was used in the pop single “Barbie Girl”. A YouTube user called PluggingAI offers songs imitating the voices of the late rappers Tupac and Notorious BIG

“An artist’s voice is often the most valuable part of their earnings and public persona, and stealing it, by any means, is wrong,” Universal Music General Counsel Jeffrey Harleston told US lawmakers last month. .

Discussions between Google and Universal Music are at an early stage, and there is no scheduled release date for the product, but the goal is to develop a tool for fans to legitimately create these tracks and pay the copyright owners for it, according to people close to the case. Artists would have the option to accept it or not, people said.

Warner Music, the world’s third-largest record label, is also in talks with Google about a product, a person familiar with the matter said.

Music executives liken the rise of AI-generated music to the early days of Google-owned YouTube, when users began adding popular songs as soundtracks to the videos they created. The music industry spent years fighting YouTube over copyright infringement, but the two sides established a system that now pays the music industry an estimated $2 billion a year for these user-generated videos.

As AI has developed, some big stars have voiced apprehension that their work will be diluted by fake versions of their songs and voices.

The issue came to prominence earlier this year when an AI-produced song mimicking the voices of Drake and The Weeknd went viral on the internet. Universal Music, the record label for Drake, Taylor Swift and other popular musicians, removed the song from streaming platforms for copyright infringement.

In April, Drake criticized another song that used AI to mimic his voice, calling it “the last straw”, while rapper Ice Cube described these cloned tracks as “demonic”.

Other artists have embraced the technology. Electronic artist Grimes has offered to let people use their voice in AI-generated songs and split the royalties.

“There are some good things,” she told Wired magazine this week, referring to the AI ​​tracks using her voice. “They’re so in line with what my new album could be that it was kind of disturbing… On the other hand, it’s like, ‘Oh, I can live forever.’ I support self-replication.”

Robert Kyncl, chief executive of Warner Music, told investors on Tuesday that “with the right framework in place” AI could “allow fans to pay their heroes the greatest tribute through a new level of of user-driven content… including new cover versions and mash-ups” [combinação de músicas diferentes].

He added that artists should have the option to approve. “Some may not like it, and that’s totally okay,” he said.

For Google, creating a music product could help it compete with rivals such as Microsoft, which has invested $10 billion in leading artificial intelligence company OpenAI, which owns the market-leading model known as GPT-4.

The model has already been integrated into Microsoft’s Bing search engine and productivity software, and Google has rushed to catch up by releasing its own AI products such as the Bard chatbot.

Universal Music asked streaming platforms in April to prevent AI services from copying its music without permission or payment, the Financial Times reported. The company, which controls about a third of the global music market, has asked Spotify and Apple to cut access to its music catalogs to developers who use them to train AI technology.

Lyor Cohen, a former record executive who heads YouTube’s music division, is working on the project for Google, according to people familiar with the matter.

In January, Google featured in an academic paper an AI-powered version of music software that could generate music from text descriptions such as “animated arcade game” or “reggaeton fused with electronic dance,” and more commands. detailed as “a quiet violin melody backed by a distorted guitar riff”.

At the time, the company said it had “no plans” to release the tool commercially, and the authors pointed to limitations, including possible copyright infringement, when the software played songs by specific artists from your training data. In May, however, Google released the experimental tool known as MusicLM to consumers and said it is working collaboratively with artists to develop it.

Google, Universal Music and Warner Music declined to comment.