AI is making journalistic language more repetitive and predictable – and it’s a problem for all of us

2026-05-27 · Source: Artificial intelligence (AI) – The Conversation · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Publishing & Journalism · Depth: Intermediate, short

Summary

The increasing use of generative AI in journalism is making public language more repetitive and predictable, diminishing its richness and capacity for innovation. Large language models (LLMs) prioritize statistical regularity, leading to formulaic expressions and a neutral tone. A critical risk is "model collapse," where AI systems train on previously AI-generated text, reducing linguistic variation, entrenching existing biases, and homogenizing writing. This process limits journalism's ability to mediate specialized registers, teach new forms of expression, and foster pragmatic nuance like irony or ambiguity. While judicious mixing of synthetic and human data can mitigate decline, the mass replacement of human writing by AI threatens the public language ecosystem and society's capacity for nuanced public debate.

Key takeaway

For editorial analysts and tech journalists integrating AI into content creation, you must critically assess its impact on linguistic diversity. Recognize that mass replacement of human writing with AI, especially when used for subsequent training, risks "model collapse" and the homogenization of public language. Prioritize judiciously mixing synthetic and human data to preserve the press's role in fostering linguistic innovation and nuanced public discourse, rather than allowing AI to diminish it.

Key insights

AI's pervasive use in journalism risks linguistic homogenization and decline, particularly via "model collapse" from training on synthetic text.

Principles

• LLMs favor statistical regularity, producing predictable text.
• AI training on synthetic text causes "model collapse."
• Linguistic variation reduction entrenches existing biases.

In practice

• Mix synthetic and human data to contain errors.
• Avoid mass replacement of human writing by AI.
• Preserve human text for linguistic diversity.

Topics

• Generative AI
• Large Language Models
• Model Collapse
• Journalism Ethics
• Linguistic Diversity
• Public Discourse

Best for: AI Scientist, Research Scientist, AI Ethicist, Policy Maker, Tech Journalist

Related on AIssential

• How AI English and human English differ – and how to decide when to use artificial language — AI English lacks human variation and readability, favoring formal, dense "exam English" due to training biases.
• The Evolution of Reasoning in Small Language Models [Yejin Choi] - 761 — Focused investment and innovative data strategies can significantly enhance small language models' reasoning capabilities and address AI homogeneity.
• On the troubling rise of generative AI suspicion in academic publishing — Overextended suspicion of GenAI in academic publishing leads to false positives and an "AI witch hunt."
• More than half of new articles on the internet are being written by AI — Over half of new online articles are AI-generated, shifting content creation towards formulaic, low-stakes writing.
• The Sequence Knowledge #862: Learning About Text Diffusion Models — Autoregressive LLMs suffer from sequential generation pathologies, suggesting text diffusion models offer a promising alternative.
• How the AI Writing Panic Is Making Us All Worse Writers — Stylistic evasion of AI writing patterns is a temporary, self-defeating strategy against advanced pattern-matching models.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial intelligence (AI) – The Conversation.