Concern is growing that generative AI will accelerate inequality in the entertainment industry—enabling superstars to be "everywhere all at once" while reducing opportunities for lesser stars and behind-the-scenes staff. Plausibly, the phenomenon will also become visible outside entertainment, with the highest performing professionals augmenting themselves with new automation tools to become highly productive—while squeezing out lower-performing workers.
In 2023, the threat of generative AI to the entertainment industry was a major driver behind actors' and writers' unions' strikes that shut down production at major Hollywood studios. Both strikes were settled in fall, in terms that generally favored the rights of actors and writers (most commentators saw the victories as conferring historic protections against AI)—but some aspects of the settlements may yet prove controversial. In particular, a provision in the SAG-AFTRA (Screen Actors Guild – American Federation of Television and Radio Artists) deal allows for the creation of synthetic performers and of digital replicas of real performers. Critics argue that this could allow big-name stars to feature in multiple projects at once, pushing out lesser-known actors. Already, AI-powered dubbing allows performers to broadcast in multiple languages in real time, and in their own voice. Tools to automatically lip sync such broadcasts are imminent. Notably, Disney has acquired the rights to the voice of actor James Earl Jones who provides the voice of Darth Vader in various Star Wars movies. Presumably, Disney plans to use this voice in productions even long after the actor's death (James Earl Jones is 92). In music, the Abba Voyage show in London uses digital avatars to create a new Abba show, and a new Elvis production (Elvis Evolution: beginning in London before moving to Las Vegas, Tokyo and Berlin) will combine AI and holograms to create "new" Elvis performances.
Overall, increased use of AI in entertainment could have the effect of providing a kind of omnipotence to superstars whilst squeezing opportunities for other performers. This threat is not lost on entertainment critics. One write up about the forthcoming Elvis show argues that audiences would be much better off supporting the current generation of struggling musicians rather than watching a hologram of a long-dead star.
Plausibly, the idea of AI giving "superpowers" to a small elite whilst squeezing out the mainstream will apply just as much to humdrum corporate work as to Hollywood. Increasingly, corporate workers including lawyers, management consultants, and administrative staff, will have access to AI tools that automate aspects of their work and make them more productive. Examples include Microsoft's CoPilot (an option with Office 365) and corporate-built chatbots such as Deloitte's PairD. Similarly, computer programmers will increasingly utilize tools that generate draft code.
Various commentators have noted that generative AI automation might have a disproportionate impact on white collar, professional occupations. A recent Goldman Sachs report estimates that 44% of legal work can be automated. Similarly, a recent UK government report estimates that AI will have the biggest disruption on employees with degree-level qualifications such as management consultants, financial managers, accountants, psychologists, economists and lawyers.
But whilst automation of factory workers and call-center staff has been implemented from the top-down, with firms implementing tools that directly automate workers, white-collar professionals are—in effect—being asked to augment themselves to become more productive. Typically, firms are providing professionals with access to generative AI tools that staff can exploit as they choose. Firms likely hope to slow hiring and/or layoff a percentage of staff—but doing so requires a form of "bottom-up" automation to occur, as workers compete with one another to become more productive. The situation seems likely to result in some employees that become hyper-productive and others that fall by the wayside.
Already, firms offer outsize rewards for star performers and performance reviews commonly involve ranking staff (including firing the lowest performing cohort—"rank and yank" or "stack ranking"). Probably, star performers will be best placed to exploit the generative tools that their firms provide. Low-performing employees may struggle to use the tools or create automations that only deliver mediocrity.
For employers, a smaller pool of hyper-productive star performers is far better than a large pool of mixed-ability workers. Arguably, this is the unstated goal of the vast majority of corporate generative-AI projects.
In the future, in fields as diverse as music and accountancy, generative AI could accelerate inequality by giving superpowers to an elite class of AI-augmented workers that eliminate the need for large numbers of mid-ranking employees. And as entertainment is already starting to demonstrate, even opportunities to become part of the "augmented elite" may decline in the long term—as workers and performers sell the rights to their synthetic replicas that then outlive them. Perhaps, in the future, AI versions of the past stars of law and accountancy will thwart the ambitions of young graduates, just as Elvis's hologram outcompetes today's emerging musicians.