Pattern index

Common AI writing patterns

Each entry includes pattern name, why it is flagged, detector context, and practical rewrite guidance. This is a diagnostic resource, not an evasion playbook.

Pattern: Meta-preface stack

Why flagged: repetitive lead-ins like "In today's world" and "It is important to note" often cluster in generated drafts.

Detector context: this pattern often co-occurs with low-burstiness sentence rhythm.

Practical rewrite guidance: cut one layer of framing and start with the concrete claim.

Pattern: Template transition ladder

Why flagged: overuse of transitions such as "Furthermore," "Moreover," and "In conclusion," creates predictable sequencing.

Detector context: lexical repetition contributes to confidence inflation.

Practical rewrite guidance: replace stock connectors with direct logical movement or paragraph breaks.

Pattern: Rhythm flattening

Why flagged: narrow sentence-length variance can indicate machine-like cadence.

Detector context: low standard deviation and low burstiness are common supporting signals.

Practical rewrite guidance: mix sentence lengths and vary clause depth where it helps clarity.

Pattern: Safe abstraction drift

Why flagged: abstract statements with few grounded specifics make writing feel generic.

Detector context: weak entity density plus stock phrasing increases risk scores.

Practical rewrite guidance: add concrete examples, named entities, and verifiable details.

Calibration links

Cross-check with evidence: Do AI Detectors Work? and the False Positive Hall of Fame.