Skip to main content

Absence of evidence

Some fields in DecAIHub project cards contain values like unknown, none, or positioning only. This page explains why such values do not mean the project lacks technology — and why they also do not mean the technology exists.

In short: DecAIHub records the presence or absence of verifiable evidence, not the presence or absence of the technology itself. If evidence is absent, it is an open question — not a verdict.

Core principle

DecAIHub operates on the principle of evidence-based assessment: a conclusion is made only on the basis of verifiable artifacts (Tier-1 and, to a lesser extent, Tier-2). If no artifacts are found, the conclusion reads:

Tier-1 confirmation not found.

This is not the same as:

The technology does not exist.

The project may:

  • have a working product but not publish its code or documentation;
  • be at an early stage where artifacts have not yet appeared;
  • use a closed development model;
  • have evidence that was not found at the time of verification.

In all of these cases, DecAIHub assigns 0 to the corresponding flag and records: "what needs to be provided or verified for this to become 1." When new artifacts appear, the card is updated.

How to read status values

unknown

Means: no information on this aspect has been found in any source (Tier-1, Tier-2, Tier-3).

  • This is not a denial: we are not stating that something does not exist.
  • This is not a confirmation: we are not filling in the gaps for the project.
  • This is a record: "as of the verification date, no verifiable data was found."

none

Means: nothing has been detected for this aspect. Used when the context implies a specific answer (e.g., "AI component: none" — no AI component identified).

Like unknown, this is an evidence-reading mode. The difference is contextual: none is more commonly used for fields where a specific value is expected.

positioning only

Means: AI is mentioned in the project's marketing or positioning, but no Tier-1 evidence of a working AI component has been found.

  • The project may use the word "AI" in its name, description, or category.
  • No code, metrics, demos, or technical descriptions confirming AI at the product level have been found.
  • This is not a verdict and not an accusation. It is a record: a claim exists, evidence — not yet.

Why DecAIHub does not fill in the gaps for projects

Conservatism is a deliberate choice. The reasons:

  • Protection against AI-washing. If claims were accepted without evidence, any project could receive a high score based on claims alone.
  • Protection of the reader. The reader should see what is substantiated and what is not — and decide for themselves.
  • Updatability. The card records the state as of a specific date; when Tier-1 artifacts appear, the assessment is revised.

A rule from the DecAIHub snapshot:

When data is insufficient or uncertain, assign 0.

This means: it is better to record a low score when data is absent than to inflate it on the basis of unverified claims.

What to do if you see unknown or none in a card

  1. Check the verification date (Last verified). The card may be outdated, and new data may have emerged since then.
  2. Review Evidence. It shows which sources were used and what is missing.
  3. Check the links. If the project has documentation, a repository, or an audit — verify independently.
  4. Do not equate absence with denial. unknown means "we do not know," not "this does not exist."

Do not confuse with

  • unknown is not 0/6. unknown is a value of a specific field; DecAI Fit = 0 means ai_score = 0 (no confirmed AI component).
  • positioning only is not a verdict. It is a gap between a claim and evidence that may be closed in the future.
  • A conservative assessment is not a negative assessment. It records the boundaries of what is verifiable right now.

See also

No comments to display

Back to top