Bengaluru: Pop music has been getting angrier and darker—and the trend shows no signs of reversing. A new study published in Scientific Reports analysed over six decades of music—with the help of AI—and rated it based on ten moral dimensions
The study, led by researchers at Queen Mary University of London, studied pop music lyrics from 1960 to 2023. Using two large datasets—the WASABI corpus of two million songs (1960–2010) and Billboard year-end charting songs (1960–2023)—the team fed lyrics through AI models trained to detect dimensions derived from Moral Foundations Theory. These are Care/Harm, Fairness/Cheating, Loyalty/Betrayal, Authority/Subversion, and Purity/Degradation.
Expressions of moral vices—Harm, Cheating, Subversion, Degradation—increased substantially, while moral virtues such as Care and Purity declined. In the Billboard data alone, vice-related foundations showed increases ranging from roughly 36 to 71 per cent over the study period, while Care and Purity each declined by over 20 per cent.
The researchers noted that the rise of genres like hip-hop introduced a lyrical discourse that critiques systemic inequality—themes that the AI models may flag as moral vices, but which represent a deliberate challenge to hegemonic norms.
The models struggled most with Fairness, a foundation that resists easy detection in lyrical language. It detected fairness in both positive and negative lyrics.
“These moral dimensions reflected in lyrics might emphasise themes of justice, power struggles, and political criticism, either as calls for fairness or as critiques of authority,” the study reads.
Also read: Mohammed Rafi to sing again. This time in an AI avatar
Gender and genre
Gender differences in the data are also notable. Lyrics by female artists tended to score higher on Care, while male and mixed-gender artists were more frequently associated with Harm. These gaps were less pronounced in mainstream pop charts, suggesting that commercial pressures produce a kind of moral homogenisation.
“Betrayal and Authority, by contrast, displayed relatively flat trajectories with limited gender differentiation…Purity showed some fluctuation, rising during the 1970 s before gradually declining, with slightly higher expression among female artists,” the study reads.
Genre is a strong predictor of moral content. Harm and Degradation were most accurately predicted within metal music, Purity was best captured in religious lyrics, and Care was most legible in R&B and soul.
The study is correlational. And the researchers caution against assigning causation.
“It remains unclear if music actively shapes societal morals or simply mirrors them, though the reality is likely a complex, bidirectional exchange,” it reads.
What the data does establish clearly is the direction of travel. Across six decades, the arc of the popular song has bent not toward justice, but toward vice.