AI writing can often be spotted by long sentences, nominalizations, and overused “and”
Afinetheorem · x · 2026-08-04
The post argues that AI-generated writing can be spotted through a recognizable set of stylistic tells, then turns those tells into a possible evaluation target for labs.
From the cited Economist-style breakdown and the image, the main signals are:
- longer sentences with less punctuation, and overuse of “and”
- frequent trios/lists, which also inflate “and” usage
- polysyllabic adjectives like “significant” and “increasingly”
- scientific jargon such as “rate-limiting” and “parameter”
- nominalizations, e.g. turning verbs into nouns like “expansion”
- the common pattern “it’s not X, it’s Y”
The image adds a more detailed writing guide: keep ideas in chronological order, use short sentences, make actors explicit, cut nominalizations, avoid excessive “to be” verbs, prefer simplicity, and reduce cosmetic errors. The core takeaway is that these rules could be turned into an eval for training or detecting AI writing.
Related event: The Economist Lists Telltale Signs of AI Writing(2 posts)→
More from Research
- Paper photo shows “Fundamental Limits of Caching” in IEEE Transactions on Information Theory — cneuralnetwork · 2026-08-04
- An ARC Prize team says its CUDA C stack is already 10x faster than the baseline — GregKamradt · 2026-08-04
- Stanford researcher uses historical tasks to correct bias in synthetic data — arena · 2026-08-04
- Wayve says its GAIA-4 world model is now useful for robotics simulation at scale — alexgkendall · 2026-08-04
- NVIDIA open-sources NOOA, a Python object-oriented framework for AI agents — Roger_M_Taylor · 2026-08-04
- Light Origins demo shows a humanoid robot switching between walking, climbing, and vaulting — Distinct-Question-16 · 2026-08-04