I’m FOR AI-generated songs being allowed on streaming platforms, but with transparency and rules. That distinction gives us a much stronger position than pretending there are no legitimate concerns. AI-generated music should belong on streaming platforms because the purpose of a streaming service is to provide listeners access to music, not to enforce a particular method of creation. P1. A song's eligibility for a streaming platform should primarily depend on whether it is legally distributable, complies with platform rules, and does not deceive listeners or unlawfully appropriate another artist's identity or work. P2. The fact that AI was used to create a song does not, by itself, establish that the song is fraudulent, harmful, or unwanted. P3. AI can be used in different ways: completely generating a track, assisting with vocals or instrumentation, helping with production, or functioning as one tool within a human-led creative process. Spotify itself now treats AI use as a spectrum rather than a simple “AI vs. human” binary and has introduced AI credits for transparency. � Spotify +1 P4. Therefore, banning all AI-generated music because some AI music is deceptive or exploitative commits a hasty generalization: it takes problems associated with some uses of the technology and applies them to every use. P5. The legitimate concerns—streaming fraud, impersonation, copyright violations, and misleading listeners—can instead be addressed through disclosure, detection, demonetization of fraudulent streams, and removal of violating content. P6. Streaming platforms are already demonstrating that distinction is possible. Deezer reports that fully AI-generated tracks accounted for more than half of its new uploads at peak levels in June 2026, while still accounting for only about 1–3% of total streams after its filtering and recommendation policies. Deezer also reports that up to 85% of streams on fully AI-generated music were detected as fraudulent and removed from royalty calculations. � Deezer Newsroom +1 C. Therefore, the existence of AI-generated music does not justify excluding it from streaming platforms as a category. The better policy question is how it should be regulated and disclosed, not whether it should exist on the platforms at all. And there's an important distinction here: allowing AI music is not the same thing as saying AI music should be treated identically to human-created music in every context. A platform can allow AI-generated songs while requiring clear labeling, preventing AI impersonation, excluding fraudulent streams from royalties, and protecting artists whose identities or copyrighted material are misused. That's already the direction some platforms are taking. Spotify has introduced AI-related credits and artist-identity protections, while Deezer labels fully AI-generated music and excludes detected AI tracks from algorithmic and editorial recommendations. � Spotify +1 So if the opposing argument is: “AI music can harm human artists.” I agree that it can. But that doesn't establish: “Therefore AI music shouldn't be on streaming platforms.” That's a non sequitur. We don't normally eliminate an entire category of technology because some people abuse it. We establish rules governing its legitimate use. And there's another important consideration: listeners themselves should have agency. If someone wants only human-created music, give them filters and clear labels. If someone enjoys AI-assisted or fully AI-generated music, they should be able to listen to it. Deezer's 2025 survey found that 80% of respondents wanted AI-generated music clearly labeled, which supports transparency as a meaningful alternative to an outright ban. � Deezer Newsroom So my position isn't “AI should replace human musicians.” It's: Let AI-generated music exist on streaming platforms, but make the system transparent, enforce anti-fraud rules, protect artists from impersonation and unauthorized use, and let listeners decide what they want to hear. The question isn't “AI or humans?” It's “How do we allow new forms of music without allowing deception and exploitation?” That is a solvable policy problem, and it doesn't require banning an entire category of music.

Your argument appeals to something real—human creativity, artist livelihoods, and the danger of flooding platforms with low-effort content—but it doesn't establish the conclusion that AI-generated songs therefore don't belong on streaming platforms. You're moving from concerns about some uses of AI to a categorical judgment about all AI-generated music. Your core argument seems to be: P1. Streaming platforms should represent human expression. P2. AI can generate enormous quantities of music without human artistic experience. P3. Large quantities of AI music can crowd out human artists. C. Therefore, AI-generated music shouldn't be on streaming platforms. The problem is that P1–P3 don't logically entail C. At most, they establish that unrestricted AI uploads could create problems. That supports regulation, not necessarily exclusion. And your statement that streaming platforms are a “contract between listeners and human expression” is an assertion that needs justification. A streaming platform isn't inherently limited to music created exclusively by humans. Technology has always changed what counts as a musical production process—digital audio workstations, synthesizers, drum machines, Auto-Tune, sampling, and other tools have all changed how music is created. Spotify itself explicitly describes AI use as a spectrum, ranging from assistance in particular parts

Your strongest point is the volume problem, and I agree that it deserves serious attention. But it still doesn't establish your conclusion that fully AI-generated songs therefore “don't belong” on streaming platforms. In fact, the evidence you're citing supports a more targeted solution than a ban. You argue: P1. AI can generate enormous quantities of music. P2. That volume can create spam, fraud, discovery problems, and potential royalty dilution. P3. Human artists can therefore be harmed. C. Fully AI-generated songs shouldn't be on streaming platforms. The problem is the jump from “AI creates a serious platform-management problem” to “AI music should be excluded.” Your own Deezer example demonstrates this distinction. Deezer reported that fully AI-generated tracks exceeded 50% of new uploads at peak levels in June 2026, but those tracks represented only about 1–3% of total streams after Deezer's detection and recommendation policies. Deezer also says it removes fraudulent streams from royalty calculations and excludes detected AI music from algorithmic and editorial recommendations. That matters because it directly challenges the inference that massive upload volume necessarily means massive displacement of human listening or royalties. You say: > “Human artists then compete against infinite supply.” But upload supply isn't the same thing as listener attention. A platform can receive 90,000 AI tracks without giving those tracks equivalent exposure. Deezer's current system is literally an example of separating those two things: enormous upload volume, but only 1–3% of total streams. And your claim that “transparency doesn't fix a flood” is only partly right. Transparency alone doesn't fix fraud, but transparency isn't the only proposed mechanism. Detection, recommendation controls, demonetization, and removal of fraudulent material can address different parts of the problem. Spotify, for example, has implemented AI disclosure through credits and has also introduced protections against AI-generated artist identities. Spotify says its AI Persona profiles are excluded by default from editorial and algorithmic recommendations, while its AI credits allow artists to disclose how AI contributed to a track. So the actual policy alternatives aren't merely: A. Allow unlimited AI music with no rules or B. Ban AI-generated music. There's a third option: C. Allow it, but regulate it. That can mean: Clearly label fully AI-generated music. Prevent AI impersonation of real artists. Remove fraudulent streams from royalty calculations. Prevent mass-generated spam from dominating recommendations. Require disclosure of significant AI involvement. Remove content that violates copyright or platform rules. Give listeners filtering or discovery controls. And importantly, I'm not arguing that AI music should be treated identically to human music. There are legitimate reasons to distinguish them. I'm arguing that distinction doesn't logically require exclusion. There's also a hidden premise in your position: > Any category capable of flooding a platform should be removed rather than regulated. Why should we accept that principle? If the problem is fraud, attack fraud. If the problem is spam, attack spam. If the problem is impersonation, attack impersonation. If the problem is royalty dilution, prevent fraudulent streams from entering the royalty pool. If the problem is listener deception, label the content. Those are targeted solutions to the actual harms you've identified. And your claim that an AI “can't mean” a melody is actually a separate philosophical question about whether AI possesses intentions or experiences. Even if I grant your conclusion that current AI lacks human subjective experience, that still doesn't establish that listeners should be prohibited from hearing music generated by it. A listener can value a song because of how it sounds, the emotions it evokes, or the creativity involved in using the technology. The listener's experience doesn't disappear merely because the production method is different. So I think you've demonstrated something important—but different from your conclusion: AI-generated music can create genuine problems for streaming ecosystems. I agree. But the existence of those problems doesn't establish: AI-generated music doesn't belong on streaming platforms. It establishes: AI-generated music requires rules appropriate to the risks it creates. That's a much more defensible conclusion—and current platform practices show that detection, labeling, recommendation controls, and anti-fraud measures can actually be implemented.