How AI is changing language

2026-07-04 · Source: AI (artificial intelligence) | The Guardian · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Content Creation & Production, Publishing & Journalism · Depth: Intermediate, long

Summary

The widespread adoption of AI-generated text is profoundly altering language, making it challenging to distinguish human from machine writing and influencing human linguistic patterns. Forensic linguist Claire Hardaker's "Bot or Not" test reveals people identify AI text correctly only about 60% of the time, often relying on unreliable cues. This uncertainty fuels paranoia in literary and media circles, leading to accusations against authors and withdrawn publications. While commercial detectors like Pangram claim high efficacy, they can be fooled. Large Language Models (LLMs) exhibit distinct patterns, such as overusing "focal words" like "delve" and "boast," favoring nouns, and homogenizing diverse English styles into an Anglo-American standard, termed "cultural ghosting." This AI-speak is now impacting human conversations. Novelists Jennifer Egan and Jeanette Winterson offer contrasting views on AI's role in creative writing, with Egan avoiding it due to "infection" fears and Winterson embracing it as a tool, both acknowledging AI's current inability to replicate human embodiment and social experience crucial for originality.

Key takeaway

For writers, editors, and content creators evaluating text authenticity, relying on intuitive "tells" or commercial AI detectors is unreliable due to AI's sophisticated mimicry and human stylistic overlap. You should recognize that AI's pervasive influence can subtly homogenize language, potentially affecting your own writing style. While LLMs can serve as functional tools for basic tasks, prioritize human embodiment and social experience to cultivate true originality and emotional depth in creative work, actively resisting linguistic flattening.

Key insights

AI's linguistic influence blurs human-machine writing distinctions and reshapes human language, challenging authenticity and originality.

Principles

• Human-like "tells" are unreliable for AI detection.
• LLMs lack embodied experience for true originality.
• AI models tend to homogenize linguistic styles.

In practice

• AI detectors are unreliable for definitive judgments.
• LLMs can assist with basic writing tasks.
• Monitor your own writing for AI stylistic influence.

Topics

• AI Text Detection
• Large Language Models
• Linguistic Stylometry
• Cultural Ghosting
• Creative Writing
• Human-AI Interaction

Best for: NLP Engineer, AI Scientist, Research Scientist, General Interest

Related on AIssential

• AI is teaching us to speak like bots, and it’s a problem — AI chatbots are subtly homogenizing human language, making our communication increasingly machine-like by nudging expression towards an average.
• Authors Guild test finds some AI detectors perfectly identify human writing while others fail on every single text — AI text detectors vary widely in accuracy, with some reliably identifying human text while others frequently produce false positives.
• How to Humanize AI Text: What We Learned from 19,804 Text Pairs — Targeted AI text rewriting can substantially lower detector scores, but its effectiveness varies by topic and measurement.
• Content for Content’s Sake — LLM-generated content is subtly altering human language and communication, eroding trust and overwhelming digital systems.
• Why Social Media Text Breaks AI Models — AI struggles with social media text due to its informality, code-mixing, context-dependency, and rapid evolution.
• Operation-Guided Progressive Human-to-AI Text Transformation Benchmark for Multi-Granularity AI-Text Detection — AI-text detectability is non-monotonic, influenced by edit operations and revision history, not just AI content proportion.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI (artificial intelligence) | The Guardian.