AI Content Audit Expert

What it addresses

AI-generated professional text can contain plausible but unreliable patent numbers, statutes, case law, and statistics. Manual review often samples only, while web cross-checking can be misled when contaminated sources repeat the same error. This skill turns audit work into explicit rules: flag high-risk facts, validate argument structure, and clean templated phrasing.

How it works

• Data validation: checks patent-number uniqueness, bibliographic consistency, statute versions, case-law status, and patent legal status. It anchors on official databases and separates blacklisted sources, whitelisted sources, and data that still needs verification.
• Logic validation: for highly law-related writing, it reviews the IRAC structure—issue, rule, application, and conclusion; for other content, it checks causal chains, claim consistency, and evidentiary support.
• Style validation: detects formulaic transitions such as “It is worth noting that” and “In conclusion,” plus vague praise, noun stacking, passive overload, and long sentences, then suggests more precise professional phrasing.
  The workflow typically covers content import, contamination filtering, module review, risk marking, and report output.

Where to use it

This skill is best suited for auditing patent, legal, case-law, and analytical texts. It does not replace factual judgment or legal advice. Recent statutory changes, unindexed judgments, cross-border patents, and niche matters should be routed to human review; high-risk content should be confirmed by qualified professionals.