It might be considered the first real big fight between humans and robots. To summarize into a very, very simple, and incomplete explanation, it happens at a time when the labor union of people who work on movies have to re-negotiate their contract with the people who pay them. They went on strike for 118 days, the longest actors’ strike in Hollywood history. Now they have a new contract.
The problem was that if there's one thing artificial intelligence can do, it's art (as opposed to law, medicine, engineering, etc.). Not saying it's necessarily good at art, but it's sure as hell good enough for the people who pay for movies to be made, and who then go on to make money off those movies. Not sure if it's good enough for the people who pay to watch these movies, but then again, it doesn't seem like anyone needs to care about consumers anymore anyway, so...the people who work on movies are in for a fight if they want to be treated like humans.
The problem, for the rest of us, is that we're humans too. Well, not if you're reading this, because the only people who read this weblog are in fact robots. But the rest of us are humans. And for the first time since Africa, we are facing real competition.
Below is a report about the 2023 SAG-AFTRA labor story, and some good bits about how humans and robots are being positioned against each other in the labor market, and what that might mean for us in the future. It's written by Data and Society Institute (citation at the very bottom, and in-text references you'll just have to get for yourself from the document; because sorry sir this is a Wendy's).
*SAG-AFTRA - Screen Actors Guild – American Federation of Television and Radio Artists
*Image credit: Gundam style mechabot made in japan - japantimes www.japantimes.co.jp
On Common Sense, Machine Learning and Big Data
Common sense, or a skewed perception of reality that perpetuates the status quo as normal, natural and unquestionable, is an expression of hegemonic power. The concept of common sense describes commonly held — yet nonetheless fragmented and heterogenous - knowledge that often goes unquestioned as fact (Gramsci,1926/1971).
According to critical theorist Hito Steyerl's (2023) power analysis of machine learning technologies, the power of the owning class depends on its seizure of data. Thus, the discussion of for-profit technology, and speci!cally various applications that track, record, and classify user data, can never be liberatory, despite the claims of the companies, industries, and institutions promising safety, accessibility, efficiency, and sustainability (Bender and Hanna, 2025; Benjamin, 2019).
The “common sense” of AI, that is, the widely-held belief that AI is a foregone conclusion, functions as an articulation of hegemonic power. In this frame, we seek to understand how the labor movement can meaningfully intervene.
AI is not simply a discursive formation that stands as a “common sense” foregone conclusion, it also obfuscates large-scale, transnational coordination of resources, labor, and people who make up the infrastructures that are required for arti!cial intelligence. This matters because it broadens the base for possible coalition building to be mobilized in these “processes of subtraction” (Steyerl, 2023, p. 12).
 |
| Data and Society, 2026 |
See Figure 1 - The four strategies on the left negotiate with the common sense of AI, giving it power and weight, as these strategies work with AI as it is, instead of disengaging with the common sense of AI, and pushing these technologies to work that are commons-based or people-centered.
Robots in Human Clothing, ie The Corporate Flesh Engine
Amazon has long referred to their Mechanical Turk platform as artificial intelligence, but there have always been humans doing the often-underpaid work of classifying and sorting content of all types, increasingly in the Global South (Crawford, 2021; Gonzalez-Cabello et al., 2025). This hidden labor has been referred to as “ghost work” (Gray and Suri, 2019; Muldoon et al., 2024) and “human-fueled automation” (Irani, 2019), drawing attention to the people who power AI systems.
Synthetic Performers (i.e., entirely digitally-produced performers created through generative AI). The [new negotiated contract] establishes guidelines around the creation and use of “digital replicas” and “synthetic performers.”
Digital replicas are digital reproductions of an actor's voice or likeness. The contract described two types of digital replicas: “employment-based digital replicas” and “independently created digital replicas.”
Employment-based digital replicas are those that are created during the actors’ physical participation in work through methods such as scanning, which can then be used to depict the actor in scenes they did not actually perform.
Independently created digital replicas are made without the actor's physical participation and can perform in scenes that they did not perform. The parameters of consent and compensation vary depending on the type of technology utilized. The guidelines for synthetic performers created through generative AI are less robust than those for digital replicas. Synthetic performers are entirely digitally-produced performers created through generative AI that do not resemble a recognizable actor and are not voiced by a person. The contract requires that studios who want to use a synthetic performer must notify and provide the union with opportunities to bargain over the usage of a synthetic performer in lieu of hiring a human performer. A document drafted by the union with frequently asked questions on AI notes that “for wholly synthetic assets, [studios] cannot use them without notifying the union and bargaining. Had we not done that, there was nothing stopping them from using these synthetic assets without anyone's consent.”
The contract also notes that if studios create a synthetic performer through prompting a generative AI system using a performer's name and their “principal facial feature” — the mouth, nose, eyes, or ears — that is recognizable, studios must bargain with the performer and obtain their consent.
However, for synthetic performers and independently created digital replicas, there are exceptions for consent with regards to uses protected by the First Amendment. A summary of the tentative agreement lists these exceptions as “comment, criticism, scholarship, satire or parody, use in a docudrama, or historical or biographical work.” In the aforementioned types of projects, studios do not need to obtain consent from performers to use their digital doubles.
--Source: Big Data & Society. Dis/engaging the ‘common sense’ of AI: Labor strategies
from the 2023 SAG-AFTRA around data-driven technologies. Mar 2 2026. Emma May, Britt Paris and Serita Sargent, Rutgers School of Communication & Information.
No comments:
Post a Comment