Unsupported Generalization

Short Explanation

Unsupported Generalization is the gap between the strength of a claim and the strength of its evidence. The reference describes it through two AI failure modes: amplification of unsupported sentiment, where a general impression in the notes becomes a finding in the record; and confidence inflation, where qualified or uncertain language becomes declarative while the factual basis does not change. It also appears across records: where similar characterizations repeat for the same employee, the volume creates an appearance of a documented pattern even though no individual entry contains an anchored incident. The reference is direct: repetition is not corroboration, and consistency in drafting inputs is not the same as consistency in fact. The remedy is to scope each claim to what the anchored evidence actually establishes.

Why It Matters

A pattern conclusion that rests on fewer than two anchored, documented incidents is an escalation trigger in the reference. Generalized claims are the easiest part of a file to challenge and the most damaging when they fail.

Reviewer Questions

• Is each pattern or evaluative claim scoped to what the anchored evidence establishes?
• Did qualified or uncertain language in the notes become declarative in the record?
• Does a pattern conclusion rest on at least two anchored, documented incidents?
• Has repetition across records been mistaken for corroboration?

Common Failure Pattern

Related JRS Sections