Overcoming AI’s ‘Storytelling’ Problem: Lessons From The Music Industry

AI can write poetry, create art, and compose music, but it still can't tell a decent story. The people building the models are researchers with scientific backgrounds who never studied the arts. That disconnect explains a growing frustration with generative AI: it lacks true narrative ability.

A new analysis from Forbes argues the music industry offers a roadmap for overcoming AI's storytelling problem. The industry weathered its own technological disruptions—sampling, digital distribution, AI composition—by embracing collaboration between engineers and artists. Today's generative AI companies, by contrast, remain siloed, dominated by mathematicians and computer scientists. The result: models that generate technically impressive but emotionally flat prose.

“The people that are building the models are researchers who went to school, but didn’t learn about the arts,” the article notes, quoting an unnamed industry observer. That observation cuts to the heart of why AI-generated novels, scripts, and marketing copy often feel hollow. Without understanding narrative beats, character arcs, or emotional pacing, models simply predict the next most probable word.

The AI storytelling problem is not new. Early text generators produced nonsensical rambles; today's large language models (LLMs) can mimic style but rarely sustain a coherent plot. OpenAI's GPT-4, Google's Gemini, and Anthropic's Claude have all been shown to lose narrative threads after a few paragraphs. A 2025 Stanford study found that 78% of AI-generated short stories had logical inconsistencies that a human reader would catch.

Music faced a similar challenge. When sampling emerged in the 1980s, purists called it theft. When AI began composing pop hits in the 2020s, critics said it lacked soul. But the music industry adapted by forming hybrid teams: producers who understood both code and chord progressions. Today, AI tools like Amper Music and AIVA are used alongside human composers.

The same principle applies to narrative AI. Companies that pair linguists, playwrights, and poets with machine-learning engineers produce better storytellers than pure-engineering teams. Startups like Sudowrite and NovelAI already employ creative writers to fine-tune prompts and training data. The result is output that readers find engaging, not sterile.

This interdisciplinary approach has broader implications. As AI-generated content floods marketing, entertainment, and journalism, the ability to tell a compelling story becomes a competitive differentiator. Brands that rely solely on algorithmic content risk losing audience trust. Human-AI collaboration, not replacement, is the winning formula.

Looking ahead, expect more tech companies to hire from the humanities. The next breakthrough in generative AI may come not from a better architecture, but from a writer in the room. The music industry showed that technology and artistry can coexist. It's time for AI storytelling to listen to that lesson.