TrustMap
Methodology

Methodology

TrustMap scores are built from only objectively observable facts. Every rule is public.

Score structure (100 points)

The total score is the weighted sum of six categories.

Category 1Professional Background

20 pts

Objective evidence behind a creator's authority on AI

  • ・AI-related work history (OpenAI / Anthropic / Google etc., AI implementation experience)
  • ・Public technical output (open-source code, peer-reviewed papers, books, talks)
  • ・Community record (Hugging Face, Kaggle, OSS contributions)
  • ・Academic / educational backing (degrees, Kaggle medals, etc.)

Category 2Source Citation Rate

15 pts

How often claims come with sources or evidence

  • ・Sample 10 posts from the last three months
  • ・Compute the share of factual claims that carry a source link
  • ・Citations to primary sources (papers, official announcements) are weighted higher

Category 3Fact-Check History

15 pts

Accuracy of past claims and the integrity of corrections

  • ・Confirmed misinformation in past posts (-5 points per major incident)
  • ・Public correction record adds points (integrity signal)
  • ・Consistency over time and clear separation of speculation from fact

Category 4Transparency & Traceability

15 pts

Verifiability after the fact and disclosure of conflicts

  • ・Major past posts remain online and traceable (≥80% retention over the last 6 months)
  • ・Promotional posts disclose conflicts (PR / sponsored / tie-up labels)
  • ・Real-name / identity disclosure (not required, but adds points)

Category 5Content Originality

10 pts

Independent thinking vs. AI-generated or recycled content

  • ・AI-generation rate measured with detection tools (GPTZero / Pangram)
  • ・Spot-check that posts are not just translations or summaries of others
  • ・Using AI is not penalised. Posting raw AI output as one's own thinking is.

Category 6Earnings & Achievement VerificationFlagship

25 pts

Evidence behind 'I made money with AI' claims, and matches to scam patterns

  • ・Verifiable evidence for earnings claims (payment dashboards, tax filings, third-party witnesses)
  • ・Transparency on revenue sources (what AI service / product produced the income)
  • ・Scam-pattern detection (high-ticket funnels, lifestyle-photo-heavy feeds, copy-paste hype templates)
  • ・Continuity / traceability (does the past business still exist?)

Total-score labels

高信頼
High Trust85-100 pts

Claims are well-supported; trust level is high

概ね信頼
Mostly Trusted65-84 pts

Generally trustworthy; some claims warrant verification

注意
Caution45-64 pts

Neutral. Review the evidence before relying on it

要検証
Needs Review25-44 pts

Claims are not sufficiently backed by evidence

信頼困難
Low Trust0-24 pts

Reliability of claims could not be established. Proceed with caution

要警戒

The flagship: Category 6 (Earnings & Achievement Verification)

For income claims like "I made $10K/month with AI," we independently evaluate the verifiable evidence behind the claim and whether it matches typical scam patterns.

When a creator's Category 6 score is 6 points or below, a special "Warning" label is shown. It is a consumer-protective cue: "Please verify carefully on your own before paying any money."

※ This is not a determination that any individual is committing fraud. Scores are evaluations based on observed facts.

Operating rules

Evaluation window
Posts from the last three months (X, note, YouTube, blogs, etc.).
Sample size
10% of posts, with a minimum of 5 and maximum of 20.
Update frequency
Monthly.
Notice to evaluated creators
We email each evaluated creator when their page goes live and share a dispute channel (in progress).
Who we evaluate
Individual creators (operating publicly under their own identity). Organisation accounts, media companies, and official corporate handles are out of scope. A title like "CEO of XYZ Inc." is fine as long as the person posts as an individual.

Disputes and correction requests

We accept "there is a factual error" or "please reflect newer information" requests from the evaluated person. Fact-based corrections are reflected promptly.

The dispute form is in preparation.

⚐ Important

  • ・Our scores are evaluations based on observable facts, not a determination that an individual is committing fraud.
  • ・The final judgment is yours. Treat scores as one input among many.
  • ・Our intent is consumer-protective information sharing, not attacks or defamation against individuals.
  • ・The site is funded entirely by paid subscriptions. We accept no sponsorship or affiliate revenue.
See pricing →← About TrustMap