Why AI Worship Music Feels Empty

What are the essential elements of worship music? My wife and I heard a stirring worship song the other day. Not only were the lyrics theologically rich, but the music was moving as well. Still, something seemed off. By the end of the song, we realized it was AI-generated—not merely written by AI, but performed by it as well. If “performed” is even the right word. My wife’s response was, “That makes me sad.”

Our sadness was the correct response, and it is revealing. It was a split-second recognition that something was not right. Our hearts were not made sad by having perfectly articulated some philosophical ideas surrounding AI worship music, which had dampened our spirits. It was the other way around. Our hearts were instinctively saddened, and we began trying to articulate why.

What was it our hearts were responding to? First, I think it had to do with how much we liked the song. If the song were subpar, we would have felt no disappointment upon learning it was artificial. We lost something because we liked it, but what did we lose?

What we lost was the connection to the song’s writer and the passionate musicians who created it. Music is communal. When a song resonates, it is because you are resonating with someone else. You know there is someone out there, whether it be the author or the performer, who understands something you understand—their hearts were moved in the creation of the song, which in turn moves us. Instinctively, we thought there was a brother or sister in Christ out there who was incredibly gifted musically and using their talents to glorify our great God. That was not so, so our hearts grew sad.

The fact that no one was behind it except the AI prompter made it feel heartless. The prompter may have had heart, but there was no heart behind the voice we heard singing. It was not a person who had suffered, believed, wrestled, repented, and rejoiced before our Savior. Which means there was no worship in the voice. It was artificial passion, artificial grit, and artificial resonance. AI can simulate worship, but it cannot worship. 

Finally, there is the aspect that great art usually comes at great cost. There was very little existential cost or effort involved in creating, performing, and producing this song.  This does two things. Firstly, as the world is flooded with this kind of “content,” true acts of creative courage and strength, even if they are not as good as the AI song, will be drowned out by lesser talents. Secondly, if some worship songs are written and performed through heartfelt inspiration and others are more artificial, cynicism will begin to spread to all of it, not just the artificial ones. Once we know some voices are hollow, we begin to wonder if all of them are. Not only will we have lost the one song when we find out it was AI-generated, but it may also thin our response to all the worship music we hear.

My wife’s sadness was a form of clarity. She was recognizing internally that worship is not only about theological soundness and moving music; it is also about who is worshipping and whether they are offering it to God.

-D. Eaton

Author’s Notes

I am still working through where I believe AI is appropriate and where I believe it is not. This short experience and writing only scratches the surface and leaves me with many more questions to explore. Here are a few.

• Would the sadness have been as profound if the song had only been written by AI but performed by someone with a heart of worship? I don’t think it would have been as sad. Many people passionately perform songs they have not written. There is still a connection to a real person. Had it been the other way around—written by a brother or sister in Christ but performed by AI—I think the sadness would still have been significant. But all this will need to be unpacked another time.
• Does this apply only to musical performances? What about AI voices reading scripts and writing? I ask this question because I often use AI voices to read my articles on YouTube.  I do believe there is something more detrimental in AI musical performances than reading, but maybe I am being hypocritical. In reading, we are primarily moved by the words themselves. We don’t lose as much when AI reads to us. In music, we are moved just as much by the performance as the words and notes—sometimes more. Unlike reading articles or stories, music has a syntax all its own that cannot be communicated by reading the notes.

What do you think?