The world of music as we knew it is changing. The advent of artificial intelligence (AI) is reshaping the way we create, comprehend, and appreciate this universal form of self-expression. Whether or not you are tech-savvy, you’ve probably come across AI-generated music, maybe unknowingly, thanks to streaming platforms like Deezer, as they strive to introspect this intersection of music and AI.
Deezer isn’t just playing around with a few algorithms though. They took this exploration a step further by commissioning a survey in partnership with research firm Ipsos. The outcome of this venture was both surprising and unsettling. With 97 percent of participants unable to distinguish AI-generated music from compositions made by human artists, the survey has definitely stirred up a serious debate. But don’t jump to conclusions just yet, as the scenario isn’t as black and white as it seems.
The experiment roped in about 9,000 participant listeners. Each participant was given three tracks to listen to and then tasked with singling out the track that was wholly AI-generated. The catch line, however, was that if a participant failed to identify all three tracks correctly, they were instantly grouped into the ‘fail’ category. This simply means if you could identify two out of the three tracks, you were still deemed unsuccessful.
The process, though rigorous, ends up looming as a tad bit unfair for the participants. If you, as a participant, correctly distinguished two tracks, you have demonstrated some accuracy in discerning AI-generated music from human-made compositions. However, Deezer and Ipsos still categorize you under the broad umbrella of those who cannot differentiate.
This exaggerated generalization fuels an important question: How fair is it to conclude just based on this study that a staggering 97 percent of people can’t differentiate AI-generated music from human-made music? The survey, though groundbreaking, seems to paint a sweeping assumption about human receptivity and discernment in face of evolving technology in music.
While the survey findings might be eye-opening, it’s evident we haven’t reached the final chapter of this fascinating discourse. There’s more to understanding, appreciating and creating music than passing an artificially constructed, boolean test. As we continue to navigate this exciting blend of technology and music, the dialogue is sure to evolve, adding new layers of depth to our shared musical experience.
For more nuanced details of the study, please refer to the full story on The Verge.
