Suno, ChatGPT and the machine-made music flooding your feed 🎼

Cliquez ici pour lire en français

Scroll through your social media feed long enough and you’ll almost certainly encounter it — that eerily polished track with no recognizable name behind it, or the song that sounds just familiar enough to hum along to without knowing why. What most people don’t realize is that a machine may have made it. Tools like ChatGPT and Claude are now being used to craft and refine song lyrics, while audio generators like Suno handle the rest — composing the melody, shaping the rhythm, and delivering a fully synthetic voice. The result carries a sonic quality that makes the boundary between human artistry and algorithmic output increasingly difficult to locate.

The AI musician has entered the studio 🎹

The trend has reportedly grown to a point where distinguishing the real from the convincing is becoming genuinely hard. The song « Suspendu » by musician Tim is said to be a case in point: according to M6 infos, the track is reportedly circulating in two separate versions online, with most listeners unable to tell which one came from a human. Production methods vary depending on what the creator starts with — an existing track they want to enhance, or a composition they’re building entirely from scratch.

How lyrics get the AI treatment ✍🏾

The process typically starts with a human writer. Once the lyrics are drafted, they’re submitted to general-purpose AI assistants like ChatGPT, Claude, or Copilot. These tools can match rhyme schemes, maintain consistent meter, and shape a chorus with the structural precision that good songwriting demands. This step is most relevant when the music doesn’t exist yet — if a track is already finished, there’s no need to write new words for it.

Inside the virtual studio 🎚️

No human voice required. For a fully original composition, the AI takes the provided lyrics and maps them to the musical style the user specifies in a prompt. It determines the gender of the virtual performer and the type of voice — contralto, mezzo, or a higher register — based on those parameters. More granular controls, like delivery pace or volume levels, let users fine-tune the final output. Suno and other music generators have staked out this territory, offering free and paid tiers for users to choose from based on their needs.

Where the revolution hits its limits ⚠️

AI-assisted output is deceptive — but it isn’t foolproof. To untrained ears, it passes easily. For those who know what to look for, telltale signs reportedly give it away, including the absence of certain technical markers in audio tracks when they’re imported into video editing software. On the economic side, this kind of content is also said to generate lower royalties on streaming platforms when its artificial origin is disclosed. « Music is an art form and it shouldn’t die. I’d rather not listen to music generated by AI, » says Céline, a passionate guitarist. Between the promise of a creative revolution and the fear that authentic musical craft could quietly disappear, the debate shows no sign of settling.

Does AI-generated music fool you when you hear it? Tell us in the comments.