The problem with Suno and AI music

由 David Pierce 主持的《The VergeCast》播客深入探讨了日益壮大的AI音乐世界，尤其关注它对我们创作、体验音乐以及通过音乐获利方式的影响。Pierce 邀请了《The Verge》的周末编辑、自称是音乐迷的 Terrence O'Brien 一起，共同剖析这项技术背后的复杂问题。 讨论从知名的AI音乐平台 Suno 开始。Pierce 强调了 Suno 所 *声称* 的——一个旨在民主化创作、降低音乐制作门槛的工具——与它 *实际* 表现之间的差异：它似乎更像是一种生成现有歌曲的相似翻唱或风格转换版本的方式。Terrence 引用了 Suno 首席执行官 Mikey Shulman 的争议性言论“人们并不享受制作音乐的过程”，Terrence 认为这暴露了对艺术创作的根本性误解。 Terrence 认为 Suno 发展成一个完整的音乐生态系统——既是创作工具又是流媒体服务——的想法是“反乌托邦的”。他反驳了关于 AI 音乐能实现个人艺术表达的说法，他认为，虽然它可能让个人表达自身经历，但真正的艺术和音乐关乎的是与他人建立共鸣和联系，而非自恋式的自我表达。他观察到，尽管 Suno 用户喜欢分享他们 *自己* 创作的内容，但似乎没有人有兴趣听 *别人* 用 AI 生成的内容。 一个被提出的重要担忧是“雪崩式”的AI生成音乐正在涌入流媒体服务。尽管目前AI音乐在实际收听量中只占极小一部分，但其巨大的数量仍可能淹没合法艺术家的作品。Pierce 和 O'Brien 讨论了一些服务（如 Deezer 和 Kobuz）如何标记 AI 内容，但像 Spotify 和 Apple Music 这样的主要平台却表现不一。尤其 Spotify 被认为在 AI 音乐兴起中拥有既得利益，因为它可能支付更低的版税，这与该平台过去在白噪音生成器问题上遇到的情况如出一辙。 Terrence 坚信 AI 音乐对艺术家构成了“生存威胁”，尤其是对“音乐产业中产阶级”而言。他解释说，尽管像碧昂丝这样的顶级艺术家可能将 AI 视为一种烦恼（因为 AI 模型经常“窃取”他们的音乐进行训练，即使是模糊的提示也能轻易模仿），但伴奏乐手、合写人以及那些通过小样（demo）赚钱的人则面临直接的收入损失。 话题随后转向一个具体的例子：在 TikTok 上疯传的《高飞电影》中歌曲《Eye to Eye》的 AI 翻唱版本。这个 AI 版本是在人类流行朋克乐队 Magnolia Park 发布了他们自己广受欢迎的翻唱版本之后出现的。这个 AI 版本在 TikTok 评论区引发了激烈争论，一些用户喜欢它，而另一些则因为它是 AI 生成的而强烈抵制，从而引发了关于“人类音乐与 AI 音乐”的争论。Terrence 认为，这个 AI 翻唱版本虽然看似无害，但它利用了人们的怀旧情绪，缺乏人类创作艺术的情感联结，最终从像 Magnolia Park 这样的人类艺术家那里转移了注意力和潜在收入。 他们总结道，随着 AI 音乐的质量提高，并变得越来越难以与人类创作区分开来，伦理和商业问题只会变得“越来越复杂”，从而迫使人们重新评估音乐的本质及其应有的价值。

The VergeCast, hosted by David Pierce, delves into the burgeoning world of AI music, specifically exploring its impact on how we create, experience, and monetize music. Pierce is joined by The Verge's weekend editor and self-proclaimed music nerd, Terrence O'Brien, to unpack the complex issues surrounding this technology. The discussion kicks off with Suno, a prominent AI music platform. Pierce highlights the disparity between what Suno *claims* to be – a tool for democratizing creation and lowering barriers to music-making – and what it *actually* appears to be in practice: a means for generating similar-sounding covers or genre transfers of existing songs. Terrence cites Suno CEO Mikey Shulman's controversial statement that "people don't enjoy making music," which Terrence believes reveals a fundamental misunderstanding of artistic creation. Terrence finds the idea of Suno becoming an entire music ecosystem – both a creation tool and a streaming service – to be "dystopian." He counters the argument that AI music enables personal artistic expression, suggesting that while it might allow individuals to articulate personal experiences, true art, and music, is about building empathy and connection with others, not narcissistic self-expression. He observes that while Suno users love to share what *they've* made, nobody seems interested in listening to *other people's* AI-generated content. A significant concern raised is the "avalanche" of AI-generated music flooding streaming services. While AI music currently accounts for a tiny percentage of actual listening, its sheer volume can bury the work of legitimate artists. Pierce and O'Brien discuss how some services like Deezer and Kobuz flag AI content, but major players like Spotify and Apple Music are inconsistent. Spotify, in particular, is seen as having a vested interest in the rise of AI music due to potentially lower royalty payouts, drawing parallels to the platform's past issues with white noise generators. Terrence firmly believes AI music poses an "existential threat" to artists, especially the "music industry middle class." He explains that while top artists like Beyoncé might view AI as an annoyance (as AI models often "steal" their music for training, easily mimicked even with vague prompts), session musicians, co-writers, and those earning from demos face a direct loss of income. The conversation then turns to a specific example: a viral AI cover on TikTok of "Eye to Eye" from *A Goofy Movie*. This AI version emerged after a human pop-punk band, Magnolia Park, released their own popular cover. The AI version sparked heated debates in TikTok comments, with some users loving it and others vehemently rejecting it because it was AI, leading to arguments about "human music versus AI music." Terrence argues that this AI cover, while seemingly innocuous, preys on nostalgia and lacks the emotional connection of human-made art, ultimately diverting attention and potential earnings from human artists like Magnolia Park. They conclude by acknowledging that as AI music improves in quality and becomes harder to distinguish from human creations, the ethical and commercial problems will only become "messier and messier," forcing a re-evaluation of what music is and how it should be valued.