America has shed more than 200,000 jobs in creative industries over the last four years, with roughly 50,000 lost in the past year alone. That makes for one of the worst stretches for media employment in modern US history outside of the recessions of 2001 and 2008. The downturn has coincided with rapid advances in generative AI—systems that can produce novels, images, music, and other digital art at scale—which raises a difficult question: how much of this decline is structural, and how much is being driven by new automation tools?
The Data: A Shrinking Creative Workforce
Employment losses have been broad-based across digital arts sectors, but the worst-hit subsector is movie and sound recording, which has lost more than 100,000 jobs—nearly one-third of the industry’s workforce—over the last several years. At points this summer, employment in the sector was lower than any time since the 2008 recession and close to a 30-year low. Written publishing has also seen steep declines, with employment down by more than 70,000. Overall media job losses have consistently hovered around 50,000 per year and appear to be accelerating rather than abating.
Why AI’s Role Is Hard to Measure
Pinning an exact share of these job losses on AI is extraordinarily difficult. Media firms often conceal the extent of AI use to avoid backlash and to protect legal positions around copyright. The effects of AI are also entangled with consolidation in Hollywood, offshoring of production, changing consumer tastes, and the longer-term erosion of advertising and subscription models. Still, the fact that nearly every digital media subsector is contracting while in-person entertainment remains healthy suggests something systemic is at work—one plausible contributor being AI increasingly embedded in production workflows.
Historical Echoes: From Napster to Streaming
The current disruption echoes earlier digital transformations. The music industry’s collapse in the era of file-sharing forced an entire business model rethink: recorded-music revenues declined and artists leaned on live performances and merchandise to survive. Video followed a different path—audiences moved from theaters to home viewing in successive waves (broadcast, cable, physical media), and each step expanded total watch time enough for the industry to adapt. The modern streaming era, however, presents a new squeeze: an explosion of free, user-generated content plus algorithmically produced material has intensified competition for attention, and streaming revenue has not made up for declines in ticket and cable revenue.
Different Industries, Different Fates
Different Industries, Different Fates
- Music: Streaming reduced per-unit revenues compared with the CD era, but live concerts and merchandise partially offset that decline. Music retained a live component that helped many artists survive.
- Film & TV: Hollywood has fewer avenues to replicate music’s compensation pivot. Movie ticket Sales are down roughly 25% from their 2019 peak, cable revenue is down about 18% from its peak, and streaming has not closed the gap. The result is sustained job losses in a sector with limited live-event substitutes.
- Publishing & News: This sector arguably faced the most complete unbundling. Newspapers once employed roughly half a million people in the U.S.; payrolls are now down nearly 85%. Magazine staff counts have fallen by about 66% since the 1990s, and book publishers have lost around 40% of jobs. The internet moved advertising and real-time news distribution to new platforms, hollowing out legacy business models.
- Production workflows: Roughly 32% of workers in arts, entertainment, and recreation report using generative AI to some extent. Invisible or partial AI use—writers consulting chatbots, artists using image tools for concept art—can scale rapidly across projects without ever being disclosed to consumers.
Unpredictable Spillovers and a Few Surprises
Technological disruption doesn’t only displace; it can also create unexpected demand. The recent 70% surge in book sales over six years—fueled in part by “BookTok” communities on short-form video platforms—shows how new media can revive legacy formats in unforeseen ways. Similarly, AI could produce spillovers that boost certain cultural forms or channels, in ways we cannot now anticipate. Corporate choices also matter: some major AI labs have largely avoided media generation (Anthropic) or have reprioritized resources away from video (OpenAI’s Sora), suggesting that market incentives and strategic focus will shape how much AI actually targets the creative sectors.
What This Means Beyond the Arts
While creative jobs are a small part of the overall economy, they may foreshadow broader patterns. The tech sector—where coding assistants and automation are more directly substitutive—saw nearly 300,000 job gains in 2022 followed by sustained job losses in subsequent years. So far, AI-driven displacement outside of narrow sectors remains limited: overall unemployment is higher than four years ago but still below averages seen in the past decade, and many white-collar fields such as legal and engineering have continued hiring. But if AI’s economic impact scales to match investor expectations, disruption could spread well beyond the arts.
A Cautious Bottom Line
We cannot conclude definitively that AI is the sole or dominant cause of the creative-class decline, but it is plausibly a significant and accelerating factor layered atop preexisting trends: consolidation, changing consumer attention, and technological unbundling. The future is uncertain. Policy choices (for example, public support for arts institutions), industry strategies, and unforeseen cultural responses—like the BookTok revival—will all influence whether this period becomes another painful but ultimately adaptive transition, or a deeper restructuring of how society produces and values creative work.
Licensed under CC BY-ND 4.0 International