Image Models Don’t Replace Artists; They Replace The Price Of Iteration

Text-to-image tools such as Midjourney and DALL·E let one person generate hundreds of concepts in the time it once took a team to produce a handful. That compresses labor demand for early-stage visual work (thumbnails, mood boards and variation sets) and forces a re-pricing of entry- and mid-level tasks. Even before AI, many design roles were projected to grow slowly. The U.S. Bureau of Labor Statistics projects 2% growth for graphic designers from 2024–2034, with most openings coming from churn rather than expansion.

The Christie’s moment still matters. In 2018, $432,500 for Edmond de Belamy signaled institutional comfort with machine-made work. The more important labor signal is downstream: when clients can generate infinite “good enough” options, they pay less for iteration and more for what resists automation, including brief-to-output translation, high-stakes taste and rights assurance.

Writing And Music: AI Automates The Draft Layer, And Humans Fight Over The Credit Layer

Large language models have made “first drafts” cheap: pitches, outlines, marketing copy variants and dialogue passes. That shifts labor away from producing words and toward managing risk through prompting, selection, fact-checking and brand alignment. The downside is wage pressure for routine writing and editing, plus more “spec work” buried inside workflows. The 2023 WGA MBA summary treats AI-generated text as not being literary material and requires disclosure when writers are given AI-generated material. It is an early attempt to keep AI from quietly devaluing credits and fees.

Music is following the same pattern. Tools like OpenAI’s MuseNet and Google’s Magenta showed that style can be modeled and recombined. The labor impact shows up when “functional music” (stingers, background beds and endless variations) becomes close to free. That means fewer paid hours for generating options and more emphasis on final selection, clearance and mixes that hold up under commercial scrutiny. The winners are likely to be specialists with a recognizable sound, as well as firms that can ship music as a scalable service.

Film And Animation: Synthetic Production Comes For The Middle Of The Pipeline

The biggest impact in film will not be “AI directors.” It will be synthetic production in the middle layers: storyboards, temp shots, background plates, voice placeholders and localization. That is where many careers and budgets sit, which is why negotiations have moved from ethics to economics. SAG-AFTRA’s 2023 TV/Theatrical agreement includes consent and compensation rules for digital replicas and synthetic performers (as summarized by Loeb & Loeb). The question is whether AI becomes a productivity tool that preserves careers, or a wedge that shifts revenue share from labor to platforms.

The Labor Economics: Creative Work Is Being Unbundled

Zoom out and the pattern is classic automation economics: tasks fragment, routine components get cheaper and bargaining power migrates to whoever controls the workflow. The IMF estimates almost 40% of jobs globally are exposed to AI, and Goldman Sachs Research estimates generative AI could expose the equivalent of 300 million full-time jobs to automation. In creative fields, that often looks less like mass replacement and more like fewer paid hours for drafts and variations, unless firms and workers renegotiate what gets credited, reviewed and compensated.

In day-to-day work, this shows up as “task unbundling.” A role that used to be priced as a bundle of concepting, drafting, iterating and polishing gets split into a fast, automated front half and a slower, human-controlled back half. That can polarize outcomes. A smaller number of senior creatives and editors capture more value because they approve, steer and own client relationships, while a larger pool competes for lower-margin production and cleanup work. Companies may call this efficiency, but it changes career ladders. If juniors do not get paid hours doing drafts, where do they learn? Leaders who ignore that question risk a future talent crunch because they optimized away the training ground.

The Ownership Fight: Prompts Don’t Equal Rights

IP still matters because it determines who gets paid. The U.S. Copyright Office has emphasized that copyright protects human authorship and that applicants should disclose and disclaim AI-generated material. Its January 2025 report on copyrightability reinforces that prompts alone are generally not enough. For labor, that means credits and rights can’t be hand-waved: if ownership gets muddy, compensation gets muddy, too.

That leads to the question executives tend to avoid: Will AI raise productivity while keeping labor’s share intact, or will it widen the gap between a few highly paid creative directors and a larger pool of lower-paid operators? If platforms control models and distribution, creators become interchangeable unless they can negotiate credits, minimums, reuse payments and disclosure.

What To Do Next: Treat AI As A Workforce Redesign, Not A Tool Rollout

For leaders, the opportunity is productivity, and the risk is accidental “deprofessionalization.” Redesign jobs so AI removes low-value effort without silently removing paid skill. Define what requires human review, what earns credit and what gets logged for reuse. Otherwise, you get faster output, lower trust and a workforce that is easier to replace.

For creators, the strategy is to move up the value chain: specialize, build trust and make your contribution legible through direction, edits, approvals and rights. In an AI-saturated market, creativity becomes abundant, but compensated creative judgment becomes scarce. If you cannot be replaced, your rates will not be either.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?