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.