Deezer Implements Sweeping Policy to Remove Fraudulent and Unstreamed AI Music as Uploads Soar.

French music streaming giant Deezer announced a significant escalation in its fight against artificial intelligence-generated content, revealing on Tuesday, July 21, that it will now remove AI tracks suspected of involvement in streaming fraud, alongside those that have not accumulated any streams for at least six months. This decisive move comes in the wake of an unprecedented surge in AI music, with fully AI recordings reaching more than 50% of new music uploads on the platform at their peak level in June 2026, a figure translating to an astonishing 90,000 AI tracks. The company’s latest policy reflects a growing industry-wide concern over content saturation, financial integrity, and the ecological footprint of digital storage, signaling a pivotal moment in the evolving landscape of digital music distribution.

The Escalating Challenge of AI-Generated Content

Deezer has been at the forefront of monitoring and responding to the explosion of AI-generated music. The platform’s journey began in January 2025 when it first unveiled plans to track and tag fully AI-generated recordings using its proprietary AI detection system. At that nascent stage, the system flagged approximately 10% of daily uploads as AI-generated, an early indicator of the technological shift underway. The initial policy stipulated that while detected AI songs would be removed from algorithmic recommendations and deemed ineligible for editorial playlisting, they would remain accessible on the site for direct fan engagement. However, the sheer volume of AI content detected and the prevalent issue of streaming manipulation necessitated a more aggressive stance. In 2025 alone, Deezer’s system identified and tagged over 13.4 million fully AI-generated tracks, illustrating the scale of the challenge.

The distinction between "fully" and "partially" AI-generated songs is crucial to understanding the scope of Deezer’s efforts. Like most current AI detection tools, Deezer’s system primarily targets recordings that are 100% generated by AI models, such as those produced by prominent platforms like Suno and Udio. It does not currently flag or penalize songs where AI is used as a tool in a broader human creative process, for instance, for mastering, mixing, or generating specific instrumental parts that are then integrated into a human-composed track. This nuance highlights the ongoing technological challenge of discerning the precise level of AI involvement in music creation and the ethical complexities that arise.

Combating Streaming Fraud: A Critical Imperative

A significant driver behind Deezer’s tightened policies is the pervasive issue of streaming fraud. The company previously reported alarming statistics, indicating that up to 85% of streams on fully AI tracks were found to be fraudulent. Such manipulation schemes typically involve bots or automated systems generating artificial streams to inflate play counts, thereby siphoning off royalty payments that would otherwise go to legitimate artists. Historically, when streaming manipulation of any kind was detected, Deezer would exclude those fraudulent streams from royalty calculations, a measure aimed at protecting the financial interests of human creators. The new policy to remove such tracks entirely represents a more robust and proactive approach to tackling this illicit activity at its source.

Streaming fraud is not merely an inconvenience; it represents a significant threat to the economic viability of the music industry. Under the predominant pro-rata royalty model, where a platform’s total revenue is pooled and then distributed based on an artist’s share of total streams, fraudulent streams dilute the earnings of legitimate artists. Even micro-payments for millions of bot-generated streams can accumulate into substantial illicit gains, at the direct expense of artists and songwriters. By actively removing these tracks, Deezer aims to purify its catalog, ensure fairer royalty distribution, and restore trust in the streaming ecosystem.

Addressing the Digital Landfill: Storage and Environmental Concerns

Beyond fraud, the sheer volume of uploaded content, particularly unlistened AI tracks, poses significant operational and environmental challenges. Aurelien Herault, Chief Innovation Officer at Deezer, articulated these concerns in a previous interview with Billboard. He emphasized that storage space is a "serious topic" for the company, stating, "This is a serious topic for us, and we already removed some content that was using too much space — because it’s not only an economic issue, it’s also an ecological one." Herault’s comments underscore a rarely discussed but increasingly relevant aspect of digital content proliferation: the environmental impact of vast data storage.

Every digital track, regardless of its listenership, consumes server space and contributes to the energy footprint of data centers. The proliferation of millions of unstreamed or fraudulent AI tracks creates what many in the industry refer to as a "digital landfill." This not only incurs unnecessary storage costs for streaming services but also contributes to global energy consumption, raising questions about the sustainability of current upload models. Deezer’s decision to remove AI tracks that haven’t been streamed for six months or more directly addresses this issue, aiming to declutter its catalog and mitigate its ecological footprint. This approach aligns with a broader industry push towards more sustainable practices in technology.

Industry Reactions and Broader Implications

Alexis Lanternier, CEO of Deezer, articulated the company’s long-term commitment to these issues, stating, "Deezer has been at the frontline of fighting fraud and reducing payment dilution related to AI music for almost two years. Now that half of all daily uploads are AI-generated tracks, we are taking additional steps to safeguard the rights of artists and songwriters, while maintaining focus on music that fans actually love." Lanternier’s call for industry alignment is a critical component of Deezer’s strategy. He added, "We have reached a pivotal moment, but we have the technology to keep track of the development, and an ambition to do what’s right. If the industry can align and come together to create shared standards, we can build a healthy and sustainable music ecosystem, where great music wins."

This call for industry-wide standards resonates deeply within the music ecosystem. Artists, record labels, and legitimate independent creators are likely to welcome Deezer’s aggressive stance. For years, concerns have been mounting over the dilution of royalty pools by non-musical content, noise tracks, and now, fraudulent AI-generated streams. Measures that ensure a fairer distribution of revenue are vital for the economic survival of artists in an increasingly competitive and technologically complex landscape.

However, the policy may also spark debate among certain segments of the AI music community and developers. While the focus is on fraudulent and unstreamed content, the broader implications for legitimate AI experimentation and artistic expression remain a point of discussion. The challenge lies in striking a balance between protecting human creators and fostering innovation in AI music. As AI tools become more sophisticated, defining what constitutes "music" or "artistry" in a legal and ethical sense will become increasingly complex.

The Future of Music Streaming and AI Regulation

Deezer’s proactive approach sets a precedent that other major streaming platforms will undoubtedly observe closely. The issues of AI content saturation, fraud, and environmental impact are universal across the digital music landscape. Spotify, Apple Music, Amazon Music, and YouTube Music all grapple with similar challenges, albeit with varying degrees of transparency regarding their internal policies and data. Should Deezer’s strategy prove effective in cleaning its catalog and improving royalty fairness, it could catalyze similar policy shifts across the industry.

The long-term implications extend beyond individual company policies. Deezer’s actions highlight the urgent need for broader industry consensus and potentially, regulatory frameworks concerning AI-generated content. Questions surrounding copyright ownership for AI-generated works, the ethical responsibility of platforms to moderate content, and the development of robust, industry-standard AI detection tools are becoming increasingly pressing. A unified approach, as advocated by Lanternier, would create a more consistent and equitable environment for all stakeholders.

Furthermore, the focus on unstreamed content may prompt a re-evaluation of the "upload everything" model that has characterized digital distribution for years. While democratizing music creation and distribution was a key benefit of the streaming era, the unchecked proliferation of low-quality, fraudulent, or simply unlistened content poses significant challenges. Platforms may increasingly lean towards curation, quality control, and stricter eligibility criteria for uploads, shifting from a purely open-access model to one that balances accessibility with sustainability and artistic integrity.

In conclusion, Deezer’s latest policy marks a significant turning point in the streaming industry’s relationship with artificial intelligence. By targeting both fraudulent and inactive AI tracks, the company is addressing critical economic, operational, and environmental concerns. This bold move underscores a commitment to safeguarding artists’ rights, ensuring fair play in royalty distribution, and fostering a healthier, more sustainable music ecosystem. As AI continues to evolve, Deezer’s actions serve as a powerful signal for the imperative of responsible innovation and collaborative industry efforts to navigate the complex future of music.