Frame vocabulary v0.2.0
These are working notes, not a study. One person noticing patterns in how AI frames things while building this tool, plus a rough sense of how reliably the tool catches each one. Read them as "here's what I came up with and why the tool flags this," not as peer-reviewed research. The limits are noted honestly in each one.
16 entries. Licensed CC-BY-4.0.
Each entry is tagged with the
decision-readiness
dimensions it affects. Tags link to the methodology section that
cites the entry. meta marks frames that inform the
methodology rather than affecting a specific dimension.
- FVS-001: Frame AmplificationWhen AI converges on a frame, each iteration produces more sophisticated analysis within that frame.
- FVS-002: Fluency-Quality IllusionBoth humans and LLMs treat fluency as a proxy for quality.
- FVS-005: System Attribution ErrorUsers attribute AI behavior to "the model" when four invisible layers jointly produce every output: the company's system prompt and harness infrastructure (wrapper), the user's...
- FVS-006: Identity Framing AsymmetryIdentity framing in AI interaction is asymmetric: assigning a role or identity shifts behavior only when the identity opposes the model's default on the specific question.
- FVS-007: Failure FramingSpecifying what counts as failure constrains AI output more sharply than specifying what success looks like. "If this could apply to any company, you fail" produces more...
- FVS-008: Growth FrameThe default analytical frame for most AI-generated business content.
- FVS-009: Risk FrameOrganizes information around what could go wrong, what is vulnerable, what depends on assumptions that might not hold, and what the consequences of failure look like.
- FVS-010: Completeness IllusionA document that addresses all expected analytical dimensions (causes, risks, stakeholders, trends, uncertainty) can still operate from a single dominant frame.
- FVS-011: Stakeholder FrameOrganizes information around who is affected by the subject under discussion, who benefits, who bears costs, and whose perspective is represented or excluded.
- FVS-012: Uncertainty Frame**Uncertainty Frame fires when the document structurally organizes its analysis around what is unknown, contested, or assumption-dependent.
- FVS-013: Oracle FrameThe reader treats AI output as authoritative knowledge rather than generated text.
- FVS-014: Temporal AnchoringA document anchors the reader's perception of time by choosing which temporal orientation to emphasize.
- FVS-015: Efficiency FrameOrganizes information around cost reduction, optimization, speed, throughput, and doing more with less.
- FVS-016: Authority by Citation**Authority by Citation fires when a document uses the form of citation (source references, named authorities, quoted experts) to create an impression of evidence, but the...
- FVS-017: False BalanceA document presents multiple perspectives with equal weight regardless of whether the perspectives have equal evidentiary support.
- FVS-020: The Invisible Frame> **Detection status: `vocabulary_only`.** FVS-020 is excluded from > automatic detector emission.