Higher Ed Musings 18.

Just Say No to MLA Format in the AI World

In the last weeks of 2024, Dr. Sarah Z faced a decision that would have bewildered scholars of previous generations. Her paper "Machine Learning's Impact on Historical Research Methodologies" was ready for submission, but her choice between PMLA and American Historical Review (AHR) wasn't based on traditional metrics like impact factor or readership. Instead, she was considering how artificial intelligence would interpret her work in the coming decades.

AHR's Chicago style would allow her to include comprehensive footnotes documenting the precise versions of the AI models she used, their training parameters, and the specific databases they were tested against. Chicago's format would enable her to clearly delineate when she was building upon existing research versus presenting novel findings, a distinction crucial for future AI systems attempting to map the evolution of ideas in her field.

While PMLA's broader readership in the humanities was appealing, Dr. Z recalled her recent experience using an AI research assistant to review literature in MLA format. The system had struggled to distinguish between original research and synthesis of existing work, leading to several hours of manual verification. "In ten years," she posted on her favorite social media platform, "when AI systems are routinely processing academic literature to generate new insights, I want them to understand exactly how I reached these conclusions and what I built upon. We're not just writing for human readers anymore."

The Challenge of Citations in the Age of AI

As artificial intelligence transforms academic work, from research assistance to content generation, academics – especially humanists – must fundamentally reconsider how we document and trace the evolution of ideas. Our citation systems, developed in an era of print media and linear knowledge transmission, may no longer serve the complex web of human and AI-generated insights that characterize modern scholarship.

MLA and Chicago styles approach citation with distinctly different philosophies. MLA embraces simplicity through parenthetical citations, while Chicago employs detailed footnotes and bibliographies. This distinction fundamentally affects how readers – and now, LLMs – understand the evolution and context of ideas: who is drawing upon what and when.

Historical Context Lost

The limitations of current citation practices become particularly evident with historical texts. Consider this MLA citation:

Darwin, Charles. On the Origin of Species. Penguin Classics, 2009.

This citation masks critical historical context: some readers wouldn't know this groundbreaking work was originally published in 1859, not 2009, nor that Darwin published multiple editions with significant revisions. The evolution of Darwin's ideas—crucial for understanding the development of evolutionary theory—is invisible under MLA. When even Darwin's revolutionary ideas become flattened into a contemporary paperback citation, we must ask: what other crucial evolutionary paths of thought are we obscuring?

Translation and Cultural Context

The implications extend to works crossing linguistic and cultural boundaries. Consider:

Dostoevsky, Fyodor. Crime and Punishment. Translated by Constance Garnett, Modern Library, 1994.

This perfectly legitimate MLA citation doesn't indicate how Garnett's interpretation might differ from other translations, nor does it reveal the original publication date or the work's publication history in both Russian and English. The rich context of how this work has been interpreted across languages and cultures remains hidden.

Contemporary Scholarship Challenges

Even with contemporary scholarship, MLA's concision creates barriers to understanding. A typical in-text citation like (Smith 45) gives readers no way to determine whether Smith is presenting original research or building on others' work. The full Works Cited entry provides only marginally more insight:

Smith, John. "Analysis of Shakespeare's Sonnets." Journal of Literary Studies, vol. 10, no. 2, 2020, pp. 45-60.

AI Processing Challenges

When processing MLA-formatted scholarship, LLMs will not always be able to establish accurate relationships between sources. MLA citations regularly lead AI systems to make incorrect assumptions about the chronology and evolution of ideas. When an LLM encounters multiple MLA citations like (Smith 45) across different papers, it may erroneously conflate distinct scholars with the same surname or fail to recognize that newer papers are citing older works rather than presenting original research.

This limitation becomes particularly problematic when AI systems attempt to construct knowledge graphs or trace the development of academic discourse. Imagine a researcher in 2030 trying to trace the development of climate change theories through academic literature. In an MLA-dominated database, the system might give equal weight to a 2025 summary article and a groundbreaking 1988 study, simply because both appear as contemporary citations. The crucial timeline of scientific discovery becomes invisible to our most powerful research tools. Without clear temporal and relational information, AI systems cannot accurately reconstruct the evolution of scientific understanding or identify pivotal moments in the development of climate change theory.

The Chicago Alternative

Chicago-style citations, with their comprehensive footnoting system, provide LLMs with richer contextual information that makes scholarly relationships clearer. When processing Chicago citations, AI systems can more easily identify the full publication history of works, distinguish between primary and secondary sources, and track how ideas have been transmitted through various scholars and translations. The detailed footnotes often include commentary about the relationship between sources, making it easier for LLMs to accurately map the genealogy of ideas.

Looking Forward

MLA maintains its position in academic writing primarily because of its accessibility and efficiency. Its format aligns well with journal requirements and print publishing constraints. Its regular updates demonstrate adaptability to emerging digital sources while maintaining consistent core principles.

However, with the emergence of AI, citation systems must evolve to address how machines process and understand scholarship. How do AI systems attribute and trace scholarship origins? Can new or revised citation methods be better legible to AI systems? Can new systems capture the increasingly non-linear nature of knowledge creation? In an era where ideas can be simultaneously generated, refined, and challenged across multiple platforms and modalities, traditional linear citation systems may no longer suffice.

Just a few weeks to go before the 2025 MLA Convention in New Orleans, I’ll make a public plea to Paula Krebs: please consider changing or sunsetting MLA format! We can continue with citation systems that prioritize space efficiency over intellectual genealogy, or we can embrace the opportunity to develop new standards that serve both human and artificial intelligence. The choice is not merely about formatting: it's about whether future generations of scholars, both human and AI, will be able to fully understand and build upon our work. The time for this conversation is not in some hypothetical future, but now, as AI systems are already processing and interpreting our scholarship in ways that will shape academic discourse for decades to come.