Critical Social Justice in the Era of Large Language Models
Technology like ChatGPT demonstrates an impressive ability to generate meaningful text. Traits that allow CSJ scholarship to be hoaxed will put it at risk of domination from AI-generated articles.
Critical Social Justice is a discipline that incorporates aspects of postmodernism, critical theories, and social justice. In some sense, wokeism is a more mainstream manifestation of the thinking of Critical Social Justice theorists within the academy. Rather than just adopting typical progressive ideas, the woke often take extreme perspectives that can appear baffling even to progressives.
This is partly thanks to the influence of radically skeptical postmodern perspectives on knowledge and truth. To understand this discipline and the origin of wokeness, I recommend the book Cynical Theories (2020) by Helen Pluckrose and James Lindsay, who participated in the Grievance Studies affair with Peter Boghossian. The team hoaxed several journals into accepting preposterous and fraudulent papers in CSJ disciplines, such as fat studies and feminist geography.
The point of the hoaxing was to demonstrate that these disciplines are unrigorous. Much of this philosophy is ideologically-motivated connections between various abstractions, which are often built on other abstractions. While some disciplines are hard to understand because they rely on mathematical or scientific knowledge as a prerequisite, CSJ scholarship is difficult to understand for the wrong reasons. For example, Judith Butler won the journal Philosophy and Literature’s Bad Writing Contest for the following passage:
The move from a structuralist account in which capital is understood to structure social relations in relatively homologous ways to a view of hegemony in which power relations are subject to repetition, convergence, and rearticulation brought the question of temporality into the thinking of structure, and marked a shift from a form of Althusserian theory that takes [End of page 354] structural totalities as theoretical objects to one in which the insights into the contingent possibility of structure inaugurate a renewed conception of hegemony as bound up with the contingent sites and strategies of the rearticulation of power.[1]
Some day artificial intelligence will be able to produce extraordinarily novel and interesting philosophical arguments. Perhaps soon after, it will be able to rapidly search through tens of thousands of scientific articles, formulate hypotheses, gather data, employ complicated statistical techniques unaided by humans, and produce scientific papers without significant errors. This level of sophistication will transform the world, hopefully for the better. It may not be necessary to have the vast majority of researchers and scientists at that point. However, at the time of writing, programs like ChatGPT are not that impressive, but they do create rather interesting conversations about various topics.
If artificial intelligence is going to eat the academy, I suspect they would start with these disciplines less tethered to reality, less empirical, and less comprehensible—the same qualities that make them at risk of being hoaxed. One early effort at creating generated scholarship was The Postmodernism Generator, written by Andrew C. Bulhak in 1996. When you click the hyperlink for the generator, it creates a text in the style of postmodernism. One article I got was called “The Meaninglessness of Narrative: Postcapitalist narrative in the works of Koons,” which contained the following text:
In the works of Pynchon, a predominant concept is the concept of dialectic
sexuality. Therefore, Buxton implies that we have to choose between postcapitalist narrative and the constructivist paradigm of reality. Several constructions concerning the role of the participant as observer may be found.“Sexual identity is unattainable,” says Derrida. In a sense, Foucault uses
the term ‘Lacanist obscurity’ to denote the economy, and thus the fatal flaw,
of neotextual class. Any number of dematerialisms concerning the capitalist
paradigm of expression exist.However, Lyotard suggests the use of cultural construction to analyse sexual
identity. The main theme of the works of Pynchon is the bridge between society
and class.But in Gravity’s Rainbow, Pynchon analyses the capitalist paradigm of
expression; in The Crying of Lot 49, however, he affirms Lacanist
obscurity. Foucault uses the term ‘the capitalist paradigm of expression’ to
denote the collapse of precapitalist sexual identity.Thus, Bataille promotes the use of postcapitalist narrative to attack
sexism. If Debordist image holds, we have to choose between the capitalist
paradigm of expression and the textual paradigm of consensus.However, the subject is contextualised into a postcapitalist narrative that
includes reality as a reality. Baudrillard uses the term ‘the capitalist
paradigm of expression’ to denote the role of the reader as poet.
I doubt this article would actually be accepted by any reputable journal in studies influenced by postmodernism, but we have come a long way since 1996. With minimal editing, I suspect that soon many articles could be produced rapidly through the aid of artificial intelligence and that these articles will likely be largely indistinguishable from actually written text. Perhaps there will be a battle between researchers producing large numbers of AI-generated articles without obvious evidence of computer generation while reviewers become exceptionally scrupulous about detecting AI-like phrasing.
Researchers who can successfully pass their generated articles as genuine research can be afforded the opportunity to only produce research if they really want. Such an arrangement in which almost everything is nonsense seems preposterous, but this characterizes a lot of academia at present. Lack of rigor makes whole disciplines questionable. Even many empirical fields are plagued by bad statistical practices, most notably p-hacking. We are in a replication crisis.
Academia has incentives that are not fully aligned with the truth. Unlike a business, departments will not run out of money if they produce nothing of much value. Even worse, Critical Social Justice studies are subsidized with government money and produce negative externalities through socially harmful rhetoric and ideas. They are worse than useless.
I suspect a shift from human-written CSJ scholarship to mostly AI-generated CSJ scholarship will go relatively smoothly. Others will be able to hoax the disciplines again, but serious political opposition might be needed to do away with these fields within universities. Those who oppose wokeness need to use their political influence to get structural changes made. Academia is in drastic need of a major overhaul; there are cracks in the ivory tower. Hopefully, large language models will aid in making these systemic issues more apparent.