CYBERSECURITYTRACKER
TRACKING
Permanent story citation

Claude Opus 5.5 uses 95% fewer em dashes, but its answers are getting longer

This page keeps the story as Cybersecurity Tracker first published it. If the tracker later corrects it, the correction appears below the original and never replaces it.

Back to newsStory 8652

As cited

Copy frozen at (site build).

ai security

Claude Opus 5.5 uses 95% fewer em dashes, but its answers are getting longer

Anthropic's Claude Opus 5.5 model demonstrates shifts in writing style compared to its predecessor, including reduced use of em dashes and shorter sentence structures. Analysis indicates the newer version employs simpler vocabulary and fewer markers typically associated with artificial intelligence (AI) generated text.

Why it matters: Organizations using Claude for content generation, customer communication, or output evaluation should assess whether these stylistic changes affect detection systems, brand voice consistency, or detection of AI-generated content in their workflows.

Source published
First seen by Cybersecurity Tracker

Source attribution

Glossary