Suno, the AI music generation platform, will begin watermarking songs created on its service as legal challenges from record labels and artists grow. The company announced new tools to mark tracks, restrict downloads, and revise community guidelines to reduce misuse.
Watermarking and copyright controls
The startup will add audio watermarking and fingerprinting to stop AI-generated songs from being uploaded to streaming platforms under false claims. It is unclear whether Suno will adopt an existing system like Google’s Synth ID or build its own. No details on timing were provided.
In a blog post, co-founder and CEO Mikey Shulman stated the tools are designed to resist tampering without altering how songs sound. “They are not meant to judge whether a song is good, meaningful, or sufficiently human,” he said. “We believe artists and platforms should decide what to disclose.”
Suno also partnered with lyrics provider Musixmatch to use its Sentinel system for copyright detection. A new download policy will block mass distribution on streaming platforms, though specifics remain undisclosed.
Legal pressure mounts
The updates come after multiple lawsuits targeted the company’s practices. In June, the Recording Industry Association of America filed a case against Suno on behalf of Universal Music Group and Sony Music Group. A German court later ruled that the platform violated copyright rules under GEMA, a government-mandated licensing agency.
In November 2025, Suno experienced a data breach, which revealed that the platform scraped YouTube, Deezer, and Genius to train its models. The incident exposed data from 55 million users and led to a class action lawsuit in Massachusetts. Plaintiffs claim the company sacrificed security for growth.
Related: Exhibit at TechCrunch Disrupt 2026 this year
The startup secured $400 million in a Series D funding round in June. The funding arrived as it expanded access, letting non-musicians create songs with simple text prompts.
While the watermarking tools may help streaming platforms identify AI-generated content, they do not settle the legal debate over training data. Courts have not yet decided whether scraping copyrighted material for AI training qualifies as fair use. For now, the company seems focused on limiting damage rather than resolving the core disputes.
Updated community guidelines now prohibit “deceptive audio presented as real” and the unauthorized use of real people’s voices or likenesses. These rules aim to prevent deepfake songs and impersonation, though enforcement remains uncertain.
Shulman described the changes as part of an effort to encourage original creation. “We want to help more people make music, not replace existing artists,” he wrote. With lawsuits ongoing and regulators paying closer attention, the company’s ability to balance innovation with compliance will shape its path forward.
As the legal battles unfold, companies in similar positions are watching closely to see how courts interpret AI training practices.
