The ten minutes of Mozart that became a law for babies

In October 1993 the journal Nature published a short paper by Frances Rauscher, Gordon Shaw and Katherine Ky at the University of California, Irvine. Thirty-six college students listened to 10 minutes of a Mozart piano sonata, a relaxation tape, or silence, then did spatial reasoning tasks such as imagining how folded and cut paper would look unfolded.

After the Mozart, scores on those tasks were 8 to 9 points higher on an IQ-style scale. The effect lasted only about 10 to 15 minutes. The study involved adults, not babies, and measured one narrow skill, not general intelligence.

Retold in newspapers and adverts, it became "Mozart makes you smarter", then "Mozart makes babies smarter". In 1998 Georgia's governor, Zell Miller, proposed giving every newborn in the state a classical music recording. A whole market of music products for babies grew up around the idea.

Other labs found the result hard to repeat. A 1999 analysis by Christopher Chabris pooled 16 studies and found an average gain of about 1.5 IQ points, limited to the paper-folding task. A 2010 meta-analysis by Pietschnig and colleagues, covering about 40 studies and more than 3,000 people, found signs that weak results had gone unpublished, and found larger effects in Rauscher's own lab than elsewhere. Its conclusion was that there is little evidence for a specific Mozart effect. Researchers have suggested more ordinary explanations, such as normal variation between test sessions, or a short lift in mood and alertness from music people enjoy. Learning to play music is a separate question, which these studies did not test.

The AI angle: the myth grew by summary. Each retelling kept the exciting part and dropped the details that mattered, such as who was tested, on what, and for how long. AI summaries can do the same thing, quickly and at scale. When you ask a chatbot to summarise a study, check that the who, how many and how long survive.

The idea behind this caseLesson: Summarise anything →

1. 1When you hear a research finding, ask three questions: who was studied, how many, and how long did the effect last.
2. 2Look for replications. One small study, however famous, is a starting point, not a fact.
3. 3When you ask AI to summarise a study, tell it to keep the sample, the method and the limits, then compare its summary with the abstract.
4. 4Watch for claims that jump from one group to another, such as adults to babies, or from one skill to "intelligence".