AIA cure for Baumol's disease?

For more than fifty years, economists accepted it as near law: creative industries will always become more expensive, faster than almost everything else. It seemed inescapable. Then machines learned to write scripts, generate images, and clone voices. So what happens to the law now? The law in question is Baumol's cost disease, named for economist William Baumol, who formalised the idea with William Bowen. The core insight is elegant and brutal: when technical progress in the arts fails to keep pace with the rest of the economy, but wages rise everywhere at roughly the same rate, you get what Baumol called "irremediable cost inflation" in the arts. Think of a string quartet. It took four musicians to play it in 1780. It still takes four today. You cannot speed it up, automate the viola, or streamline the cellist. Wages have risen, but productivity hasn't budged. That's the disease. Film and television are infected in exactly the same way. Gillian Doyle and Sabine Baumann, in their 2024 paper examining artificial intelligence and Baumol's cost disease, document how media production hinges on skilled creative labour that cannot readily be mechanised or replicated. Directing, writing, performing, and editing with genuine artistic judgment — these are performing arts-style tasks where you cannot simply substitute capital for human input. So costs rise persistently, above the broader economic rate. The disease has been shaping cultural economics for more than half a century. Enter generative artificial intelligence. Doyle and Baumann survey what these tools are actually doing across media industries right now. Large language models draft scripts and copy. Image and audio synthesis generates avatars, automates dubbing, and handles lip-syncing. Post-production tools tackle editing, rotoscoping, and colour grading — tasks that once took days now take a few hours. In newsrooms, automated reports cover finance, sports, and local news as a matter of routine. Face-swap de-aging and photoreal digital doubles appear in real productions. One researcher, Connock, called artificial intelligence "creative rocket fuel" — not a replacement engine, but something that amplifies what human workers can do. That word — augment — turns out to be the key word. Doyle and Baumann find that artificial intelligence consistently augments human creative labour rather than substituting it. Humans remain in the loop, as Zysman and Nitzberg put it, to ensure coherence, quality, and ethical grounding. Caswell's research finds human editorial judgment remains the most valued skill in newsrooms. An International Monetary Fund study by Cazzaniga and colleagues predicts generative artificial intelligence will affect 40 percent of jobs worldwide — which signals massive disruption, but disruption is not the same as replacement. So does artificial intelligence cure Baumol's disease? Doyle and Baumann's answer is no. Artificial intelligence trims costs at the margins, reshapes workflows, and removes some repetitive burdens. But core creative work stays labour-intensive. The structural reliance on scarce human judgment remains. Add the policy complications — copyright disputes, licensing battles, concerns about misinformation and market dominance by major tech players — and the picture becomes even more tangled. The disease, Doyle and Baumann conclude, is still with us. Artificial intelligence is a powerful treatment. It is not a cure. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.