AI is now able to predict whether a song will be hit or not.

Why does one song get famous overnight in the music charts while another one flops? A recent study suggests that the secret to differentiating a hit exists in listeners' brain. Artificial intelligence (AI) is capable of determining this key by analysing physiological signals. Some hit song science experts aren't quite ready to congratulate themselves right now.

Researchers from Claremont Graduate University monitored heartbeats and responses of a music-listening person using a wearable smartwatch like device. They used an algorithm to convert these data into what they believe is a proxy for brain activity.

The monitor focused on a few specific emotional responses. A machine learning model was 97 percent accurate in classifying whether a song was a hit or a flop after being tested on this data. The trial results of Recent Advances in Artificial Intelligence were revealed earlier in this month.

The hit song mystery has continued for decades. Researchers designed some automated tools such as machine learning software. They can detect if this will become popular or not before a song is released. The most recent and best performing effort to address this issue is the current study.

Some experts believe that this technology might reduce the cost of producing music and build public playlists. It may eliminate the need for judges in TV talent shows. Due to its accuracy in predicting song popularity, the new method offers the potential of updating both the creative process for musicians and the distribution process for streaming services.

The study also raises questions about the reliability and moral implications of integrating brain data with artificial intelligence. It might be innovation if the study could be easily verified and generalized.

Hoda Khalil is a data scientist at Carleton University in Ontario she has done a study on the subject of hit song science but she is not involved in the current study. she believes that a machine learning experiment can be impacted by a number of biases particularly if it is trying to predict human preferences.

Even if we have sufficient statistical evidence to draw assumptions we still need to consider the potential misuse of this model. Technology cannot advance faster than ethical concerns.

In 2011 machine-learning engineers at the University of Bristol in England designed the hit potential equation. In order to examine 23 musical elements and forecast a song's popularity. The equation might be used to classify a hit with a 60% accuracy rate.

Hoda Khalil and her colleagues who analyzed data from more than 600,000 songs found no significant correlation between the number of weeks a song spent on the Billboard Hot 100 or Spotify Top 50 lists and any of the musical characteristics they evaluated.

Then researchers investigated the studies of Mike McCready the businessman who first used the term hit song science and realized at that time there wasn't enough data to support his assertion.