January 3, 2024

As generative AI has seeped into many corners of our lives, it threatens to upend the economy as we know it, from education to the film industry. Developers claim the technology will have sweeping impacts that will transform work, creating opportunities for workers and unleashing new levels of creative expression. But as generative AI comes into more widespread use, workers’ value has increasingly become synonymous with the data they create, or that which can be extracted from their labor. While the tech industry would have us believe that this transformation is inevitable, it is in no way a done deal, or even a certainty. Generative AI remains a very new technology, and there are a host of questions about its capabilities and their implications. And as we saw earlier with robotics and self-driving cars, when it comes to generative AI and labor there is excessive focus on long-term existential questions rather than the ones right in front of us. Right now, writers are seeing their work stolen wholesale by web scrapers without acknowledgement or compensation. Call center workers are being guided by AI-generated, automated responses that are an amalgam of the responses workers produced themselves in the past. In a recent photo from a fashion show, an Asian model’s face was edited with AI to make her appear white, robbing her of the exposure that is so critical to her livelihood (not to mention raising thorny issues of race and representation).

Having worked in the labor movement for over a decade, I know this is not the first time workers have been confronted with the commodification of their labor. What is also not new is the narrative that the fate of workers is sealed, that they are Luddites railing against an unstoppable machine, and therefore, that their voices are unworthy of attention. This current discourse purposely sucks the oxygen out of the room, leaving little space for discussions about what is actually happening and why it matters. But there are specific questions we should be asking about generative AI’s impacts on workers, ones that could provide us with meaningful, constructive ways to disrupt the current trajectory. Here are some of them:

• • How is generative AI scaffolded on previous systems and structures, and how might that be reinforcing historic power dynamics? Generative AI did not emerge from thin air: existing economic models, social norms, and even political conditions supported its rapid adoption. Examining these conditions could reveal the technology’s uneven effects across different industries, labor markets, and worker populations.

• • If there is a new “AI economy,” how might it follow patterns we recognize from the gig economy? Many of the industries now facing disruption by generative AI technologies have precarious, non-standard employment relationships, similar to workers in the “gig economy.” While concerns about generative AI were front and center in the Hollywood writers and actors strike, for example, a larger looming concern was the rise of streaming services and the potential for minimal residuals, shortened production schedules, and lean writing rooms. As the shift to streaming restructures the entertainment industry, it is increasingly characterized by short-term, discrete, often customer-centric modes of service. Enabling studios to use generative AI in production adds another layer of precarity to the entire industry. What can workers’ experiences in other gig economy jobs tell us about what impact generative AI might have?

• • What equity concerns are raised by the use of generative AI in different fields? Societal biases and embedded hierarchies about who and what type of work is valuable will inform how generative AI is rolled out and who benefits from it. For example, some artists of color have expressed concerns about the economic and cultural implications of AI-generated art based on their work, including appropriation. When technology determines what culture gets represented and whose views of the world are advanced, what are the larger costs?

• • How can we recognize the value of work — beyond the form it takes? Actors, writers, and models are finding that generative AI can reduce them to their image, their words on a page, and even their measurements. Moreover, there are growing examples of false or flawed AI generated work. How does this emerging technology diminish the value of workers and their contributions, and how might we recognize it differently? Does data accurately capture the knowledge and skill that humans bring to their production? What is lost or appropriated in the process of extraction, and how might it be regained or reconstituted?

• • How is AI legitimized and how does this impact workers? In the past, when automation threatened to displace jobs, the pro-automation argument has been that some jobs are undesirable and technological innovation could create new (if not necessarily better) ones. The reality is that automation has more often led to a hollowing out of the middle class, resulting in underemployment or malemployment. In cases where some but not all work was automated, that work became more poorly paid, difficult, or precarious for workers — like cashiers who were left manning multiple self-checkout registers. With generative AI being used to produce the kind of “creative” work that once seemed safe from automation, who is this function serving? Are there any benefits to be found for workers and artists?

• • How are workers actually relating to AI technologies? Workers are responding in a myriad of ways to generative AI. Some are being managed by AI systems while others are using it to augment their work. Still others are being forced to use these systems, to integrate it into their daily work as a way of keeping up with a particular machine or the industry as a whole. Others have lost their jobs, while AI has led to the creation of new ones (often, with little clarity about who can access those jobs). Still others resist such technologies, sabotaging or hacking them to meet their needs. What do these varied experiences tell us about whether AI will truly benefit or harm humanity?

• • The prior questions inevitably lead us to this one: What are our options and potential solutions? Are we limited to strengthening copyright laws and reskilling? How else can workers regain agency in the workplace?

No one person has the answers to these questions. But it’s also true that not everyone is being given the space to respond, or present questions of their own. Workers who are subjected to and use these tools need to be a part of the conversation about what will preserve their livelihoods and creativity, and in some cases their identity and privacy. Their perspectives can show us what questions and considerations are missing from the public discourse.

To help create that space and tackle some of these critical questions, Data & Society’s Labor Futures program is holding a series of public talks. Starting with a January 18 conversation about hierarchy, in these interdisciplinary discussions we will uncover how new technological systems are impacting worker agency and power. Join us as we dissect generative AI’s benefits and harms, counter hype, and expand our collective knowledge and power. For more information and to RSVP, please visit our website.