2024 January 29
I spent most of 2023 on three overlapping pursuits: finishing drafts of two books, writing music for three albums, and learning to work with artificial intelligence.
At the same time, I posted regularly on X (Twitter) and elsewhere about AI’s practical value to artists and businesses. Keeping up with the field felt like drinking from a firehose; major advances arrived every few days, and what had been reliable last week was often obsolete by the next.
Over the holidays the distance between those of us who work with these tools every day and everyone else became impossible to ignore. For people inside the field, AI is already a working instrument, and we look ahead to the moment when it can convincingly simulate consciousness–an event that will eventually touch nearly every life. Outside that circle, the conversation is dominated by questions, apprehension, and deep philosophical resistance to the idea that hardware and software might one day possess anything resembling intelligence, let alone consciousness.
Business
The AI systems I rely on for business planning, operations, content, books, music, images, voice-overs, and video are built on large language models. For text and operations I use ChatGPT and Grok; for images, Adobe Firefly and Midjourney; for video, Runway; for voice, ElevenLabs.
One clear business example is video advertising. Last year we moved from the traditional sequence (concept, script, voice-over, music, casting, live shoot, edit, color grade) to a hybrid workflow:
-
- Concept (human, with AI as a sounding board)
- Script (human, with AI as a sounding board)
- Voice-over (AI)
- Music (AI)
- Images and video (AI)
- Editing and color grading (human)
What once required four to six weeks now takes about five days, with cost reductions, depending on complexity, of 50 to 90 percent.
Art
In music the time and cost savings are equally dramatic, yet legal constraints remain substantial. Using AI to generate a song for anything beyond advertising is still problematic, though that is likely to change. My forecast is that within five years a listener will be able to tell a sound system what mood they are in and receive a custom playlist of AI-generated music tailored to that moment. The same shift will reach film and television: a spoken request such as “Make me a sixty-minute science-fiction romantic comedy” will produce and stream a personalized movie on demand.
As a composer I have taken a page from Brian Eno’s approach to synthesizers: encouraging the machine to generate results one would not have imagined. I maintain hundreds of pages of prompts designed to push music AI toward unexpected “hallucinations.” I then slice those results in a music editor, turn them into patches for my own synthesizers, and process everything through layers of effects.
I’ve also used text-based AI as a sounding board for the story behind some of the music I’ve written in 2023. The key is to not use AI to just spit out a story or lyrics, but to train the AI as a persona of a half a dozen people who you’d like to have on your soundboard and then have a conversation with it. From that conversation, then write your own story and lyrics. The advantage is that any combination of voices I want is available at any hour.
I applied this process to two songs for my Ankh project, “Little Lies” and “Whispers in the Night.” For comparison, I wrote two other songs without AI input: “Miss Understood” for Null Paradox and “Homewrecker” for a new artist.
Resistance
The resistance I encounter tends to fall into two categories. The first is that no machine can possess the heart and soul of a human being, so any art it produces will remain emotionally distant from listeners. This is ultimately a philosophical question each person must settle for themselves. When the MP3 format arrived, many of us dismissed it because the sound quality was poor; I was among them. Most listeners eventually accepted the trade-off for convenience, while a smaller group continued to seek higher-resolution files (I’m in this group) and some returned to vinyl.
Something similar may occur with AI art: broad acceptance alongside durable niches for work made entirely by humans.
The second objection (AI is displacing human jobs) is already visible. In video advertising I went from collaborating with many people to collaborating with none in a matter of months. For anyone whose livelihood depends on writing copy, recording voice-overs, or shooting live action, the practical choices are to master these tools, to focus on the areas where human judgment and presence remain difficult to replicate, or to move into adjacent fields. Another path is to claim the space that will always exist for work made by humans for humans; the contemporary equivalent of the vinyl revival or the continued demand for skilled portrait painters even after photography became ubiquitous.
The larger development is more encouraging. Just as the integration of good cameras into phones allowed far more people to tell visual stories, widespread access to capable AI is lowering the barrier to creating music, video, images, and writing. We’re entering what may prove to be a golden age of storytelling for everyone who wants to participate.
READY TO EXPLORE THE FULL PROCESS?
Join our community on Patreon for complete articles, creative process breakdowns, minimalist frameworks, and downloadable tools — including reusable templates and behind-the-scenes access.