Graphite study reveals thousands of new AI writing tells
A new study by Graphite researchers has identified thousands of previously unknown "tells" in AI-generated writing, indicating that current AI models still leave detectable patterns. These findings suggest that despite advancements, AI-written content can often be distinguished from human-written text through specific linguistic markers. The study highlights examples like "this matters" as a common phrase used by models such as Opus 5.5.
3 articles from 3 outlets covered this story. The underlying claim is sourced from a paper.
What do all outlets agree on?
3 outlets covered “Graphite study reveals thousands of new AI writing tells”. All of them report the following:
- Researchers identified new AI writing tells
- Thousands of AI writing tells were found
- AI models still leave detectable patterns
- "This matters" is an example of an AI tell
- Opus 5.5 is an AI model exhibiting these tells
- The findings come from a study by Graphite
Which outlets covered this?
All 3 articles found on this story, grouped by the stance of the piece. Every link goes to the original publisher.