MBW Views presents a series of op-eds featuring insights from influential figures in the music industry. This piece is authored by Krystle Delgado, a prominent Entertainment and IP attorney who leads Delgado Entertainment Law PLLC and hosts the widely followed Top Music Attorney YouTube Podcast, boasting over 200, 000 subscribers. Delgado is notably involved in high-profile class-action lawsuits against generative AI music platforms SUNO and UDIO. In this op-ed, Delgado addresses the ongoing debate around “Walled Gardens” versus “Open Studios” in AI music. She notes that the music industry has been evasive about whether AI companies using artists’ work without permission is a serious issue or just an awkward phase before formal agreements are reached. The discourse has shifted from accusing these companies of illegally taking millions of songs to a narrative suggesting mutual profit, primarily benefiting major labels. However, many of these songs belong to independent artists who were never asked for consent, compensated, or included in the new AI licensing agreements. The potential for independent artists to gain justice through ongoing lawsuits against Suno and Udio remains uncertain. The industry is moving toward models like the “walled garden, ” which allows users to create and share music within a closed platform without exporting AI-generated music externally.

Major players like Warner, Universal, and Merlin have agreed to this structure with Udio to ensure “properly licensed” AI models. Yet, the transparency about what music is included in these licensed pools is lacking, and independent artists have had no input. Warner has uniquely settled with Suno, which opposes the walled garden concept, allowing Warner’s catalog to be used across platforms with artist opt-in terms. This dichotomy poses a critical question: is the push for “starting fresh” truly about protecting music, or just safeguarding major label catalogs while unauthorized use of other artists’ work continues? Delgado highlights that none of these licensing efforts erase the original copyright infringements by AI companies, which were built on mass unauthorized use of music. Additionally, there's concern about whether major label artists can genuinely refuse participation in AI training despite these deals. A new complication arises for AI users, who often wrongly assume legal risks are covered by the platforms. In reality, users usually bear legal responsibility if AI-generated music leads to copyright issues. For instance, Udio has claimed in court that users—not the company—are liable for any infringements in AI-generated songs. Thus, the current AI music ecosystem involves artists having their music used without permission, users facing legal risks, and AI companies, which built their models on free artist content, remaining shielded. While AI music is undeniably here to stay, Delgado argues that the current focus on “walled gardens” misses the core problem: can AI companies genuinely “start fresh” when their foundations rest on mass piracy and theft? — Music Business Worldwide