Em-ergence of the em-dash: a population-level rise in em-dash frequency in medRxiv preprints at the dawn of the large-language-model era
arXiv:2606.29540v1 Announce Type: cross Abstract: Large language models (LLMs) can leave subtle stylistic traces in assisted text; one of the most cited is the em-dash (Unicode U+2014). Yet no one has measured whether em-dash use has changed in the scientific literature. This study, pre-registered on the Open Science Framework (HFT8C), used the full set of medRxiv full-text XML preprints from the official Text-and-Data-Mining resource. The primary cohort was first, original versions deposited 2020-2025 with an extractable Discussion section of at least 500 characters (N = 69,632). The primary
The proliferation of Large Language Models (LLMs) has led to their increasing use in text generation, making it timely to study their detectable stylistic influences on academic writing.
This study provides empirical evidence of subtle, population-level stylistic changes in scientific literature driven by AI assistance, offering insights into the evolving landscape of academic authorship and integrity.
The detectability of AI-assisted text generation in academic preprints is becoming quantifiable, moving beyond anecdotal observations to measurable stylistic shifts.
- · Linguists
- · AI ethicists
- · Academic integrity platforms
- · Authors relying solely on undetectable AI assistance
- · Reviewers unaware of AI stylistic tells
Increased focus on stylistic markers to identify AI-generated or AI-assisted content in academic submissions.
Development of new tools and methodologies for detecting AI influence in various forms of written content.
A potential shift in academic publishing policies and peer review processes to address the challenges of AI-assisted authorship.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI