NYTimes compares weekend human songwriting to AI generated tracks

In short: The New York Times published a listening challenge that puts new songs made by five humans up against AI generated music and asks readers if they can tell.

What happened

The New York Times ran an experiment published on October 3, 2026. A journalist and four experienced musicians spent a weekend in a recording studio writing and producing songs largely from scratch.

Their finished tracks were presented alongside music generated by AI. AI generated music is music made by software that can create a song after being given prompts, like instructions for what it should sound like.

Readers were invited to listen and judge for themselves. The setup makes the comparison easy to hear, but it is not the same as a controlled scientific test.

Why it matters

Tools that can generate music are becoming more common, and they raise everyday questions for listeners and creators. If AI made songs start to sound closer to human made songs, it could affect how music is marketed, how artists get paid, and how people decide what feels worth their attention.

At the same time, this kind of weekend challenge can only answer a narrow question: what this audience thought about these specific tracks, made under these specific conditions. Without published results like final vote totals or identification rates, it would be misleading to claim that listeners could, or could not, reliably tell the difference.

Source: NYTimes