Classical music 不是静止的：从 Mozart、Beethoven 到 Haydn，作曲家一直把新技术纳入创作。18世纪的自动演奏乐器曾让欧洲精英着迷；而今天的 AI 由 George E Lewis 的 Voyager 之类实验性系统延续这种传统，只是它不演奏既有作品，而是在与人实时对话中“做决定”。

与 Voyager 这类实验系统不同，ChatGPT、DALL-E 以及 Suno、Udio、Stability AI 等生成式模型以文本提示大规模生产内容。文中给出几个关键数字：OpenAI 在 2026 年估值 8520 亿美元；Unesco 预测到 2028 年，生成式 AI 可能使创作者收入减少高达 24%；Suno 上一轮融资募集 2.5 亿美元。

文章核心担忧是“skill death”：AI 可能让人们不再学习原本需要长期训练的作曲能力。PRS for Music 的调查显示，93% 的成员希望 AI 训练使用作品时创作者能获得补偿，92% 希望训练数据更透明。尽管像 Harry Yeff、Stephanie Dinkins 和 Robin Coops 看到 AI 的扩展性，但作者认为，古典音乐真正的“民主化”更可能来自教育与乐器普及，而不是大模型。

Classical music has long evolved alongside technology, from Mozart, Beethoven and Haydn writing for mechanical instruments to George E Lewis’s Voyager, an AI system from the 1980s that improvises live with humans. Voyager is experimental and non-deterministic: it “makes decisions,” so no two performances are the same, unlike historical self-playing organs that merely executed written music. (Key numbers: 18)

The article contrasts such systems with today’s generative models such as ChatGPT, DALL-E, Suno, Udio and Stability AI, which can produce huge amounts of content from prompts. Key numbers frame the scale of the change: OpenAI is valued at $852bn in 2026; Unesco projects generative AI could cut creators’ revenue by up to 24% by 2028; and Suno raised $250mn in its last funding round.

The central warning is “skill death”: if AI can generate complex music, people may stop learning the craft. PRS for Music found 93% of members want compensation when their music is used for AI training, and 92% want more transparency about training data. While artists like Harry Yeff, Stephanie Dinkins and Robin Coops see creative uses, the piece argues that regulation, not hype, will determine whether AI augments classical music or replaces the skills that sustain it.