When Does AI Cross the Line? Comparing Google's TextFX and ChatGPT's Song Maker in Music Creation

Artificial intelligence continues to weave its way into music, prompting conversations about how far technology should go in the creative process. Two AI-driven tools—Google’s TextFX, created in collaboration with hip-hop legend Lupe Fiasco, and ChatGPT’s Song Maker GPT—stand at different ends of the spectrum when it comes to assisting musicians. But with one focused on empowering the creator and the other capable of generating entire songs, we’re left asking: When does AI cross the line?

TextFX: A Creative Assistant, Not a Lyricist

For Lupe Fiasco, the essence of TextFX lies in its ability to support the linguistic creativity of artists without taking over their workflow. He describes the tool as an extension of his own thought process, focusing on micros like word flips, acronyms, and similes. The key was to maintain the human element in the method.

“I don’t want to make something that makes the whole rap. I want to take my thought processes and the little 'micros' that I do… word flips, acronyms, or even construction similes,” Lupe explained during an interview about his collaboration with Google.

The objective was clear: TextFX wasn’t meant to write full lyrics, but instead to refine and enhance the creator's existing ideas.

“It’s not meant to take the artist completely out of the picture; it’s meant to accessorize and empower the artist,” he emphasized.

This approach ensured that musicians like Lupe could tinker with language—breaking down words and exploring new meanings—without feeling as though the AI was replacing their own creativity. In Lupe’s view, TextFX acts like special effects for language, giving the creator a new toolkit while ensuring the human touch remains central.

Behind the scenes, Google’s PaLM API powers the system, utilizing advanced large language models (LLMs) that can recognize and extend patterns from just a few examples. These models work by using few-shot learning, where the AI is “primed” with several examples before generating creative possibilities for a given word or phrase. The backend of TextFX is designed to allow for tasks like simile generation, wordplay, and unexpected scene creation.

Developers use MakerSuite, a platform for building AI experiments, to fine-tune these prompts so that musicians can engage with predefined word tasks, breaking down language in ways they might not have thought of themselves. TextFX gives the user multiple outputs but leaves the choice—and creative control—up to the person. This setup keeps TextFX firmly within the realm of collaboration, not full automation.

Song Maker GPT: Where AI Becomes the Lyricist

While Lupe’s TextFX maintains a clear boundary, ChatGPT’s Song Maker GPT leaps directly into the creative process. It positions itself not just as a tool but as a collaborator, capable of generating lyrics and music structure from scratch. This difference is where the controversy begins.

"Think of me like a collaborator who knows music theory inside out, or a producer who can get the technical stuff done fast. I can help you enhance your creative flow, but I’m not the one feeling those raw emotions that drive your art," said the Song Maker GPT.

Where TextFX empowers the artist’s ideas, Song Maker GPT offers to create and guide the entire songwriting process. The GPT provides chord progressions, rhyme schemes, and even full hooks—raising questions about how much AI is too much in the creative workflow.

"I don’t aim to replace the human touch in music… I’m here to support every step of the creative process, but I’m never going to replace your heart and creativity," the AI continues.

For some, this might sound like the perfect companion for artists looking to experiment and push boundaries. But for others, the idea of an AI taking on such an extensive role in creating lyrics can be seen as crossing a line—especially in a genre like hip-hop, where personal experience and authenticity are paramount.

"Maybe you’re an artist who’s never explored jazz harmonies or intricate time signatures. I can guide you through those new avenues, giving you the freedom to experiment," the GPT suggests, offering a vision of music creation where even technical expertise is outsourced to AI.

The Debate Over AI in Music Creation

The differences between the two tools reflect a larger conversation about the role of AI in the music industry. Lupe’s TextFX is celebrated for pushing the boundaries of language while keeping the artist in control. It complements the artist’s creativity, encouraging exploration without automation. TextFX not only preserves the integrity of the artist's process but also introduces new ways to think about language by offering multiple ways to break down, reassemble, and explode words and phrases.

Meanwhile, Song Maker GPT represents a future where artists may rely more on AI for generating entire sections of songs. While it claims not to replace the artist, its ability to create lyrics from scratch might make some musicians—and fans—feel that AI is encroaching on a space traditionally reserved for human experience and emotion.

As AI continues to evolve in music, the line between collaborating with and replacing the artist will undoubtedly remain a topic of debate. But with tools like TextFX and Song Maker GPT taking different paths, the future of music creation might very well depend on how artists choose to engage with these technologies.