AI Tone Terms (English) — Reference

This document lists the English vocabulary patterns most associated with AI-generated academic prose, with a recommended per-document occurrence budget. The companion file tone-thresholds.yaml is the authoritative source consumed by deai_check.py.

How thresholds are enforced

• deai_check.py reads tone-thresholds.yaml at startup.
• Each word in term_thresholds: triggers a [Script] LOW trace once its per-document count exceeds the listed value.
• Counts are case-insensitive and word-boundary matched against visible prose (citations, refs, math, comments are stripped first).
• Override by editing the YAML; this MD file is documentation only.

Maintenance cadence (this list is a snapshot, not a final state)

This word list captures current AI-tone tells, not a permanent truth. As words such as delve and pivotal get widely named, careful authors filter them and their frequency drops, while new AI-preferred words keep emerging. Re-check this list roughly every 6 months against excess-vocabulary research and prune or add accordingly — do not treat it as frozen.

• Last reviewed: 2026-06
• Sources: Kobak et al., Sci. Adv. 2025; Geng & Trotta 2025

High-frequency AI vocabulary

These words are not banned. They are useful when used sparingly. The threshold is the point at which a reviewer is likely to flag the writing as templated.

Word	Threshold	Why it matters
significant	5	Often hides missing effect size or p-value
comprehensive	3	Marketing language; rarely earned by a single study
effective	5	Cheap claim without baseline comparison
novel	4	Reviewers discount the word unless the novelty is named
robust	4	Needs the perturbation / noise level that justifies the claim
important	5	Replace with what is at stake
various	5	Vague quantifier; usually fixable with a number
several	5	Vague quantifier
numerous	3	Vague quantifier; almost always replaceable with a count
furthermore	3	Padding connector; often signals a content-free addition
moreover	3	Padding connector
notably	3	"Notably" is rarely needed when the content is genuinely notable
remarkable	3	Editorial language; let the data carry the claim
remarkably	3	Same as above
obvious	3	Over-confident hedge
obviously	3	Over-confident hedge
clearly	4	Over-confident hedge

Burstiness (paragraph opening repetition)

When three or more consecutive paragraphs begin with the same two opening tokens, the script emits a burstiness trace. Typical offenders:

• "We propose ..." / "We propose ..." / "We propose ..."
• "In this ..." / "In this ..." / "In this ..."
• "Furthermore, ..." / "Furthermore, ..." / "Furthermore, ..."

The remedy is to rewrite at least one opener with a different syntactic shape (subordinate clause, prepositional phrase, contrastive connector).

Throat-clearing phrases

Phrases that occupy the first sentence of a paragraph without delivering information. The default pattern set covers:

• In order to better ...
• In this section, we ...
• It is worth noting that ...
• It should be noted that ...
• As mentioned earlier ...
• Leading discourse markers: Notably,, Furthermore,, Moreover,, In summary,, To summarize,

Each trigger is a single [Script] LOW trace pointing at the offending line.

Punctuation patterns

• More than max_em_dashes_per_doc em-dashes (--- or —) across the document → one aggregate trace at the first occurrence.
• Any ! in body sections (abstract through conclusion) → one trace per occurrence. Inline code, math, and comments are excluded.

Out of scope

The following are intentionally NOT enforced here:

• Sentence-level grammar (handled by analyze_grammar.py).
• Citation density (handled by verify_bib.py).
• Section structure (handled by check_format.py).
• Domain-specific terminology, which lives in forbidden-terms.md.