Discover more from The Workbench
Every Technological Change Is an Opportunity for the Owners to Rob You: Hollywood Edition
Streaming and artificial intelligence were not destined to be threats to actors and writers. The problem is that Hollywood's workers don't control how that technology is deployed.
Broadly speaking, the actors’ and writers’ strikes that have ground Hollywood to a halt are about two complaints.
First, there’s the damage that streaming did to compensation. Residuals—the pay actors and writers receive when the stuff they’ve already worked on is rewatched—almost totally collapsed in the wake of streaming platforms’ arrival. Seasons have also gotten shorter and production crews tighter, further cutting into writers’ pay and employment prospects.
Second is the rise of artificial intelligence, which could allow studios to replace a lot of human labor: write scripts without actual human writers, scan background actors once and then endlessly reproduce them in scenes after scene. All of which threatens, again, to reduce employment prospects for both groups.
Noah Millman recently argued that these are really two fundamentally different issues: The first fight is a classic fight between owners and workers over pay. The initial contracts established by the streaming platforms were just wildly unbalanced and unfair, cutting labor out of huge shares of revenue. And the strikes can and should rebalance the scales. But the second fight over A.I. is a more difficult conundrum; a classic case of new technology replacing labor, and forcing a larger social adjustment. Some forms of work have just genuinely been rendered unnecessary by A.I.—same as automobiles replacing horses and buggies—and we’re not going back.
I think Millman’s right, so far as it goes. But at a broader, forest-not-the-trees level, I think both fights are about the same basic thing. They’re both about technological shocks to a particular community’s norms of production.
More specifically, they’re about how, when those technology shocks came, one class of people within the Hollywood community—namely, the owners—basically rolled another class in the community—its workers and laborers. The owners used the shocks as an opportunity to change how the community’s revenue gets distributed internally; siphoning a much bigger cut into their own pockets, and leaving the workers with less.
The question I want to ask here is: Why did it happen like this in the first place? And why does this seem to happen so often when new technology arrives?
How the Technology Shocks of Streaming and Artificial Intelligence Were Felt
In plain English, Hollywood is a community of human beings who have all gotten together to produce a particular thing: movies and television. That community brings in revenue when it sells its product to the wider world, and it’s established contracts and agreements dictating what cut of that revenue everyone in the community will get. Crucially, those agreements and contracts were based on certain assumptions about how that production would happen—in the concrete, material sense—and how the final product would be distributed to consumers. Literally, who would physically do what, how it would get done, et cetera.
What any technological change does is it alters those concrete, material ways production and distribution happen. When that happens, the assumptions must change as well. All the agreements about who does what, and how they’ll be compensated, have to be rethought.
I think this is pretty intuitive and easy to see with A.I., again because workers are literally being replaced by technology. They’re not needed in that particular part of the production process anymore. So will they be able to find another place within the Hollywood community’s process to work? Will they have to move to a different industry entirely? What new skills will they have to learn? Will they be able to? Will there be enough work? Will those production processes have a place for them?
If all the answers are “no,” how do we as a society care for these displaced workers and the people who depended on their income? Will we?
The way the technological shock of streaming upended the community’s production and distribution process was more subtle, but just as important in its own way.
Under the old system, movies and shows were produced by one TV network or movie studio, then sold to other TV networks for rebroadcast. The networks themselves often made their money from advertisers; sometimes from subscription fees. So you had multiple players all paying one another: studios and networks paying workers for labor, networks paying studios and other networks for content, advertisers paying networks for views.
What everyone in this chain of production and distribution paid each other—including what actors and writers got paid for residuals—was based on viewership numbers. So those numbers had to be public knowledge, shared between all the players involved. That was why, if you worked on a hit TV show or movie, you could rely on a decent income from the residuals whenever that show or movie got re-shown and rebroadcast.
On top of all that, the old technology of the networks imposed certain limitations: content had to be broadcast at certain times, and couldn’t just be called up by viewers and consumed whenever. That had a physical effect on production as well, influencing the length of shows’ seasons, how many episodes ran, and how much work could be expected for how many people. In other words, the state of technology at any given time creates material realities that guide the rhythms of production, as surely as the weather and the seasons guide the rhythms of farming.
Then along comes streaming technology. Suddenly, customers can call up whatever content they want and when, while just paying a flat fee for everything. So viewership numbers become a black box known only to the streamers themselves. And all the assumptions and habits about how many episodes get made for a show, how production gets scheduled, how many writers you need and who works on what when—it all gets thrown up into the air. So all the agreements and contracts about who gets paid how much, and for what, have to be renegotiated. When that renegotiation happened, the owners and the studios managed to strike new bargains that massively reduced the cuts of revenue actors and writers were getting relative to the old agreements.
Technological Change in an Alternate Democratic Universe of Worker Ownership
I’ve made the point before that all economic activity is inescapably political. Every industry (including Hollywood) and every firm in an industry (including all the studios and production houses and streaming services in Hollywood) are just groups of human beings cooperating to accomplish a common project. To do that, they have to figure out how to get along and how to organize their efforts. And “politics” is just the process of how we human beings order our lives together.
Now think about the question of “ownership” in this context. What it really means is governorship. If you own the A.I. system that scans actors and reproduces their likeness, or that straight-up creates digital actors and screenplays, then you govern that A.I. system—you get to decide what uses it’s put to, what products it creates. And you get the money that consumers pay in exchange for that product.
Similarly, when you “own” a company (like a production studio or streamer) what it really means is you govern that company—you govern that community of human beings. You get to tell everyone in it how production will happen, how the equipment and resources belonging to the company will be used, who will do what, what happens to the revenue the company brings in, how everyone will be compensated, who gets hired and who gets fired and when and why. And if they don’t like it, then too bad.
In both cases—whether we’re talking about a piece of equipment like an A.I. system or an actual human community—to be an owner is to have power over “the means of production,” in the Marxist parlance. Crucially, this is a legal right; a product of government and law. The owner can call upon the government to use its monopoly on legitimate violence to defend her power—her governorship—over that equipment and community should anyone defy her.
You could imagine a different way of doing it. Every company of multiple people in our economy, no matter how small, is a legal entity—and the law is something our society chooses, and chooses how to design. So it could be the law that every company has to essentially be a partnership of workers: every employee is an equal co-owner, and gets an equal vote to elect the board or executives or whomever does the governing of that company. We could insist that our economic communities govern themselves the same way our actual political communities—local, state, and federal—govern themselves: as democracies.
In that alternative democratic universe, what would the arrival of streaming and A.I. have looked like?
If the actors and writers and everyone who produced content for Netflix—or any other streamer—were all worker-owners of the company with an equal voice, it seems extremely unlikely pay and residuals would’ve collapsed the way they did. Production and distribution processes would’ve changed; work schedules would’ve changed, too. But the impetus would’ve been to keep workers as whole as possible through those adjustments. Sure, workers would still vie with each other within their respective company-communities over who gets what cut, the same way blocks of voters vie over laws and policies in our current political communities, local state and federal. But the point is, everyone would’ve had an equal vote—and equal representation—in these deliberations.
Indeed, in this alternate universe, “owners” wouldn’t even exist as a separate class within the community, vying for their own interests. It would just be different groups of workers and the management they’ve elected, all jostling with each other. By definition, the bargaining would’ve turned out differently.
As for A.I. systems, which are already the kind of physical capital usually owned by firms rather than by individuals, worker-owned production companies would still have put A.I. to use. Workers want more productivity and efficiency as much as anyone, especially if they’re all joint owners of the company. More productivity and efficiency means less work, yet more revenue. But in this alternative democratic universe, the way A.I. is brought into the production process would focus, again, on keeping all the workers in the community whole and employed—on supplementing existing labor as much as possible rather than just flippantly replacing it.
The Government Makes the Law, the Law Makes the Market, Then the Market Adapts to the Technology
The point here is that technology is not inherently scary or destructive, or inherently a threat to workers and jobs—despite technology often being portrayed that way in the discourse. Owners are a threat to workers and their jobs.
More precisely, the threat comes from how we’ve designed the laws of property and ownership and market structure, so that businesses are generally hierarchical and dictatorial communities, rather than democratic ones.
Owners, of course, also like technological advances because they bring in more revenue for less cost. But the way the “less cost” part plays out is that owners are quite able and happy to kick any worker to the curb if they judge that worker no longer necessary—with no institutional obligation to hear that workers’ voice or consider their needs. And if a technological change offers the chance to redesign contracts in a way that leaves the owners with more and workers with less, you better believe the owners will jump at that opportunity.
This state of affairs gives owners a huge in-built advantage whenever it comes time to fight over pay and compensation and work conditions. Owners literally control the social fabric workers have to enter into in order to have a job and earn a wage at all. On top of that, because governing companies means controlling their revenue streams, and because this is how our society has worked for a very long time, owners as a population have built up huge stores of wealth compared to workers. If a fight breaks out between owners and workers, and halts the flow of revenue coming in, owners can afford to wait a lot longer. That’s why one Hollywood studio executive told Deadline that, “The endgame is to allow things to drag on until union members start losing their apartments and losing their houses.”
In this world, technological change is simply an opportunity owners and capitalists can exploit to rob everyday workers blind.
This brings us to the traditional function of unions: to give workers some organizational clout; to inject at least some democracy back into the mix, and bring the owners’ and workers’ bargaining position at least a little closer to equity. But all this also just pushes the scales a bit back towards balance within a system that’s just designed to massively favor the owners from the outset. Hollywood is actually more fortunate than most industries here, because they do still have active unions and a tradition of organized labor. Across most of the U.S. economy, union power has nearly been obliterated. But even for Hollywood, negotiating fair contracts and compensation agreements is largely a Sisyphean task, rolling the boulder up the hill only to see it roll back down again.
Hollywood workers have engaged in repeated strikes over the last century to secure a more just cut of the revenue when technology changed: when television was invented, when the television networks got up and running, when foreign distribution markets became a big deal, when home video rentals and purchases became a thing, and so on. Now it’s happening again. And the reason it keeps happening this way is that we, as a society, designed it to: we chose to have an economy where companies are hierarchical authoritarian systems, and where workers are always put at a profound disadvantage relative to owners.
The rhetoric of “the market” obscures these realities. That rhetoric suggests that how companies (and physical capital) are owned, and how they are governed, is just the impersonal result of natural forces that predate human society. Which is of course nonsense. No market can exist without rules—of property, of ownership, of firm governance—and rules must be chosen by society.
The government makes the law, and the law makes the market, and only then does the market “decide” all these questions—including how new technology will be adopted and what the consequences will be.
A Democratic Economy Requires Democracy Between Companies as Well as Within Them
Let’s return one last time to that alternate democratic universe: The picture I painted was one where each worker-owned Hollywood company is a group of actors, writers, and all other manner of production crew, all moving as a unit from project to project. Which is admittedly not how Hollywood actually works. Instead, actors and writers and everyone else bounces around interchangeably from project to project. Most likely, it would often work the same in our alternate democratic universe. You could imagine roving, worker-owned companies of different varieties, coming together to produce and distribute one particular show or film.
But what that means is, in order for us to have a democratic economy, in which new technology is always absorbed in a just and humane way, we must have democracy between companies as well as within them. We must have democracy at the macro level, across the whole design of our economy.
Technologies that blatantly replace labor—like artificial intelligence—again make this particularly clear. Even a democratic, worker-owned company cannot resist the practical realities of a new technology that makes it more profitable for the workers to do a particular task with automation rather than human labor. Even in our alternate democratic universe, workers would routinely find themselves displaced by new tech, and have to find a new line of work.
So how do we deal with that?
First, insist on macroeconomic policy from our federal government—taxing, spending, and interest rates—that always prioritizes full employment, keeping unemployment as low as possible and labor markets as tight as possible. (Lobbying for precisely this has traditionally been another role for unions and organized labor.) If demand for labor always outstrips supply of labor, then every worker will always have bargaining power, even when they transition to a new industry requiring new skills. We’ll need a generous welfare state, with cash support, unemployment insurance, universal health care, universal housing, well-funded education for all, and so on. That will support workers in their transitions, and see to it that no worker is disciplined into taking a job by the threat of poverty. And we’ll need a national infrastructure to help people retrain and shift careers and industries when they need to.1
Of course, the taxes to finance all this will cut down on income and wealth inequality, lowering the ceiling on how powerful any individual or company can become, even as we raise the floor. That will help insure a democratic balance of power across the whole economy.
Next, beyond question of government spending and taxing, we’ll need more specific rules for how markets are structured, and what terms companies may and may not interact with each other on.
Consider antitrust law: Matt Stoller had a helpful piece arguing that a big cause of Hollywood’s current mess is the vertical integration that came along with the adoption of streaming. That’s when every step of production and distribution has been combined in a single company. (Think of how Netflix both makes content and distributes it, and you just pay a flat fee at the tail end.) Vertical integration actually used to be illegal, until the interpretation and enforcement of antitrust law was drastically loosened in the Reagan era. And crucially, vertical integration actually destroys markets, which only exist between firms, not within them. Stoller makes a compelling case that this loss of market dynamics and price signals lead to the wildly self-destructive bets studios made in the streaming era, as well as the collapse of residuals for actors and writers—recall that residuals were calculated from viewership numbers that had to be shared between firms paying each other for different steps in the production process.
If today, we still enforced antitrust law the way we enforced it in the mid 20th-century, the arrival of streaming and A.I. would’ve played out very differently in Hollywood. And in our alternate democratic universe, well-enforced antitrust law would allow any group of workers to break off and form a new company, if they thought their fellow workers within a company were treating them unfairly. Then they could bargain firm to firm, rather than worker to worker within the firm.
The government makes the law, then the law makes the market.
Finally, there’s intellectual property, which comes up again with artificial intelligence. As I said, ownership rights are just a set of rules designed by society. They’re meant to approximate how our moral intuitions layer atop a hopelessly complex real world, and they will inevitably do this imperfectly. That’s particularly easy to see with intellectual property, which deals with abstract ideas rather than concrete possessions. And it’s doubly easy to see with A.I., because A.I. software systems literally have to be trained on huge collections of pre-existing text and images and data—all of which were created at some point by someone’s labor. And that person is not compensated by the A.I. owner.
Now, if an actor is scanned and their likeness replicated, it seems fairly straightforward they should have intellectual property rights to that likeness. But what of an actor or script generated wholesale by artificial intelligence, and thus built atop the unsung labor of countless others? Millman has a clever solution to this conundrum: all material generated by A.I.—scripts, images, etc—should be public domain. I.E. no one should have intellectual ownership over it.
Again, that’s a rule our society would have to choose, and our government would have to pass and enforce. There’s no “natural” answer to the question of how property rights should be designed.
The last, deeper truth here, is that all technology—indeed, all wealth creation, period—is like this: a collective, society-wide achievement, built by thousands of small steps and breakthroughs and insights by thousands of people over time. And only a handful of those people will wind up getting the ownership rights that actually get them a cut of the income stream that technology generates.
This has always struck me as one of the most compelling moral arguments for a universal basic income.2 We will never be able to adequately parse out each contribution an individual has made to our society’s collective wealth creation. The idea that any system of property rights, however practically necessary, can achieve that parsing is a fantasy. So the best way to handle the problem is to just guarantee every member of our society a basic stream of income, taken out of our collective wealth—wealth generated through the technology that we are all both privileged and burdened to live with.
Arguably, in a genuinely egalitarian and democratic society, “a national infrastructure to help people retrain and shift careers and industries” is what higher education would be: a lifelong offer, open to every person for free, to come in when they need to and learn something new. It’s also worth noting that in a sufficiently tight labor market, companies will be much more willing to foot the bill for training and re-skilling new workers that come in from other industries.
Of course, it’s also a compelling moral argument for making employment itself a universal human right, and for ensuring that the gap between the highest- and lowest-paying jobs doesn’t become too large. A job is, itself, one of the ways our society divvies out income from our collective wealth creation.