Confidence Adjustment

NIKOLAI proposes Confidence Adjustment as a recorded move that raises a stated risk level, or treats a capability threshold as reached, because of residual uncertainty in the underlying assessment -- with a required rationale field capturing what specific uncertainty (new incident evidence, elicitation limits, unquantified margin) motivated the adjustment. This is a NIKOLAI editorial proposal; the concept recurs across sources under different labels ('safety margin', 'abundance of caution', 'conservative upper bound') that NIKOLAI is unifying, not restating any one source's own defined term.

This is CASRAI's own proposed definition, not a definition any named organisation has agreed to. See what NIKOLAI is and is not.

Source of record

Where this definition comes from

Anthropic Risk Report, August 2026, §2.1, §2.19
“Alignment risk "explicitly raised from 'very low' to reflect increased uncertainty after 'recent incident disclosures related to model behavior in cybersecurity evaluations'" (§2.1, §2.19).”
https://www-cdn.anthropic.com/f61d49fa5596956a5dec75fea0e973bf6a6a8378/Redacted%20Risk%20Report%20August%202026%20.pdf
Gemini 3.7 Flash FSF Report, p.4
“"Safety margin: takes into account different sources of uncertainty, including the limits of our threat modeling and evaluations" (p.4); "cannot rule out being at the T/CCL".”
https://storage.googleapis.com/deepmind-media/gemini/gemini_3-7_flash_fsf_report.pdf

Crosswalk

How named organisations use this concept

Every row below is a shadow mapping. A shadow row is CASRAI's own reading of a published document. No lab, evaluator or regulator named on a shadow row has declared, endorsed, or been consulted on it. That changes only when an organisation files its own Mapping Declaration.

Organisation	Their term, as published	Match & verification	Source
AnthropicShadow mapping Anthropic Risk Report, August 2026	“Alignment risk "explicitly raised from 'very low' to reflect increased uncertainty after 'recent incident disclosures related to model behavior in cybersecurity evaluations'" (§2.1, §2.19).”	exactEQ confidence: high	Anthropic Risk Report, August 2026
OpenAIShadow mapping OpenAI Frontier Governance Framework / Preparedness Framework v2	“"out of an abundance of caution we have treated models as crossing a capability threshold in circumstances where we are unable to rule out that a new threshold had been reached, even in the absence of direct evidence that it has occurred" (FGF §2.3). Sandbagging response: "use a conservative upper bound of the model's non-sandbagged evaluation results" (PF Table 2).”	closeCL confidence: high	OpenAI Frontier Governance Framework
Google DeepMindShadow mapping Gemini 3.7 Flash FSF Report	“"Safety margin: takes into account different sources of uncertainty, including the limits of our threat modeling and evaluations" (p.4); "cannot rule out being at the T/CCL".”	exactEQ confidence: high	Gemini 3.7 Flash FSF Report
xAIShadow mapping xAI Frontier AI Framework, 30 Jun 2026	“"incorporating a margin of security" (s.2.3); not quantified.” This row cites xAI's Frontier AI Framework (30 Jun 2026, {FAIF26}), whose PDF metadata /Title reads "Privileged/Confidential DRAFT working FRAMEWORK DOC" with no xAI statement found disambiguating draft vs. final status. Treat as citing a document of unconfirmed draft/final status.	noneDU confidence: medium	xAI Frontier AI Framework, 30 Jun 2026
MetaShadow mapping Meta Advanced AI Scaling Framework v2	“"We conduct risk assessments and assign risk thresholds with maximum elicitation in mind, capturing the upper bound of risk" (§2.2.1); "provisionally rated 'high'" (§4.2.1).”	closeCL confidence: high	Meta Advanced AI Scaling Framework v2
EUShadow mapping EU GPAI Code of Practice, Safety and Security Chapter	“Measure 4.1: acceptance determination "incorporat[es] a safety margin" that must "(1) be appropriate for the systemic risk; and (2) take into account potential limitations, changes, and uncertainties of: (a) systemic risk sources (e.g. capability improvements after the time of assessment); (b) systemic risk assessments (e.g. under-elicitation of model evaluations or historical accuracy of similar assessments); and (c) the effectiveness of safety and security mitigations" -- the only source that itemises what a safety margin must account for, matching xAI's chapter-derived "margin of security" language (fn.1-2).” This row's source citation also references xAI's Frontier AI Framework, 30 Jun 2026 ({FAIF26}), for the correlated 'margin of security' language it echoes. That xAI document's PDF metadata /Title reads "Privileged/Confidential DRAFT working FRAMEWORK DOC" with no xAI statement found disambiguating draft vs. final status; the EU citation itself (EUSSC, the official 43-page chapter) is not affected, but any inference drawn about xAI's own framework from this correlation should carry the same draft-status caveat.	exactEQ confidence: high	EU GPAI Code of Practice, Safety and Security Chapter

What do these codes mean?

exact: The source term is equivalent to this element
close: The source term is close but not equivalent to this element
broad: The source term is broader than this element
narrow: The source term is narrower than this element
none: No mapping claim — used for false-friend and declared-but-undefined rows
EQ: Equivalent
CL: Close
BR: Source is broader than the element
NR: Source is narrower than the element
FF: False friend — same or similar label, different meaning
DU: Declared but undefined by the source
UV: Unverified