We need AI labels on creative content, but not for reasons you think

In March, a picture of an aggressively fashionable Pope Francis created by artificial intelligence went viral on social media.

In April, the song "Heart on My Sleeve," featuring vocals filtered to sound like Drake and The Weeknd, gained immense online popularity.

In May, movie producers expressed increasing interest in using AI to generate scripts, raising red flags for striking screenwriters.

And just last week, Microsoft called for government agencies to develop labels to indicate the role of AI in literature, visual design and music.

We do need labels for AI-generated content; however, the need goes beyond the concerns expressed by tech executives.

While the private sector is primarily interested in deepfakes and copyright infringement, we believe content labels are necessary to protect the creative community and ensure the preservation of the human-to-human connection that lies at the heart of art and entertainment.

Anxieties over algorithms competing with, or even surpassing, human creativity are nothing new. In the 1950s, the computer pioneer Alan Turing introduced the Turing test as a basic yardstick for evaluating AI's ability to generate content that resembles human efforts. The test challenges evaluators to determine whether a text or an artwork was made by a human or a machine.

If an AI vocal filter convinces you Drake sang "Heart on My Sleeve," then that AI "passed" the Turing test.

But what happens when we can't tell whether something was made by a human or a computer?

It turns out that crucial to our enjoyment of art are human empathy and communication. We watch movies, listen to music and read novels for so many reasons: captivating visuals, beautiful sounds and thrilling plot twists. Yet, these are merely means to an end — to sharing the human experience and fostering a sense of belonging to a larger community.

Throughout history, critics have argued that artists establish connections with audiences by infusing their work with their own emotions and thoughts. A love song resonates more deeply when we believe it was written by someone who has experienced love.

As algorithmic creation becomes increasingly sophisticated and as we grapple with the challenge of distinguishing between human-made and AI-generated content, the invisible human factor becomes paramount. Will we laugh at jokes or tear up at songs written by AI the same way as those crafted by a human? Imagine feeling chills while listening to a tune, only to discover that it was generated by a digital entity. How will you react? Will you emotionally bond with a computer program?

This rapidly expanding sea of Turing-proof art will create a whirlpool of uncertainty. Amid the confusion, some consumers will seek out islands of verifiably human-made content. We must remember that human programmers, institutions, training data and their inherent biases are always present, even when the content is ultimately generated by a computer model.

Luckily, we have long-established traditions of standardized labeling for content that holds value for consumers. In the U.S., the Food and Drug Administration uses standardized labeling for organic food production, and the Federal Trade Commission oversees manufacturing designations such as "Made in the USA."

In the realm of music, some combination of regulators, legislators and record labels could establish standards that define the relationship between AI and humans in the creative process. Did the singer-songwriter create all the musical and performed materials? Did a human use AI for inspiration and then construct the final product? Or was the entire product synthesized by artificial intelligence?

While AI-generated art is here to stay, and its innovations will inevitably affect jobs and revenue, we do not anticipate Turing-proof algorithms rendering human production obsolete. We believe the future cultural landscape will likely include a combination of large-scale AI generation alongside "artisanal" human production — the local bakeries of art and entertainment.

By fostering well-balanced connections between the demand for human creations and the supply of human-made content, we can ensure that, even in a world flooded with cheap and convincing AI-generated materials, we can consciously invest our time and money in art that truly evokes the human connection and empathy we deeply value.

This essay is 100% human-made content.

Chris White is assistant professor of music theory at the University of Massachusetts at Amherst and is the author of "The Music in the Data." Mariusz Kozak is Associate Professor of Music and Director of Cognitive Science at Columbia University. He is the author of "Enacting Musical Time: The Bodily Experience of New Music."