Residual Risk and Risk Acceptance Determination

NIKOLAI proposes Residual Risk Determination as a record of the assessed risk remaining after mitigations have been applied, paired with the recorded decision (and decision-maker) that this residual risk is acceptable for a stated scope -- further development, internal deployment, or external deployment. This is a NIKOLAI editorial proposal added to the candidate list (ALIGNMENT-MATRIX.md §5) because every lab framework in the corpus has some form of this determination, making it one of the most consistently-present concepts across sources, even though the exact process and who signs off varies.

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

Google DeepMind Frontier Safety Framework v3.1, glossary
“"Residual Risk Assessments: the process of evaluating the level of risk that remains after all planned mitigation strategies have been implemented." (glossary)”
https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3-1.pdf
EU GPAI Code of Practice, Safety and Security Chapter, Commitment 4, Measure 4.1
“Commitment 4, in full: "Signatories commit to specifying systemic risk acceptance criteria and determining whether the systemic risks stemming from the model are acceptable ... Signatories commit to deciding whether or not to proceed ... based on the systemic risk acceptance determination." (Measure 4.1)”
https://ec.europa.eu/newsroom/dae/redirection/document/118119
OpenAI Frontier Governance Framework, §2.5
“"If residual risks associated with the model exceed acceptable risk levels, the model is not deployed unless additional mitigation measures are implemented that sufficiently minimize risk" (§2.5); "We document a justification for why the systemic risks stemming from the model are acceptable" (§2.5).”
https://cdn.openai.com/pdf/e37d949b-8c9f-4d76-b99e-4272f4631a7e/openai-frontier-governance-framework.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 Advanced AI Framework / Anthropic Risk Report, August 2026	“AAF risk report must include an assessment of residual risk "that remains after accounting for the safety mitigations" (p.6). Risk Report bottom line: "these net out to our continued AI development and deployment to date passing a societal cost-benefit test" (§5.4).”	closeCL confidence: high	Anthropic Advanced AI Framework
OpenAIShadow mapping OpenAI Preparedness Framework v2 / Frontier Governance Framework	“"OpenAI Leadership ... is responsible for: Making all final decisions, including accepting any residual risks and making deployment go/no-go decisions" (App. B). "If residual risks associated with the model exceed acceptable risk levels, the model is not deployed unless additional mitigation measures are implemented that sufficiently minimize risk" (FGF §2.5); "We document a justification for why the systemic risks stemming from the model are acceptable" (FGF §2.5).”	exactEQ confidence: high	OpenAI Preparedness Framework v2
Google DeepMindShadow mapping Google DeepMind Frontier Safety Framework v3.1	“"Residual Risk Assessments: the process of evaluating the level of risk that remains after all planned mitigation strategies have been implemented." (glossary). Risk acceptance determination is defined in the body (s.1.3.5) with conditions per T/CCL.”	exactEQ confidence: high	Google DeepMind Frontier Safety Framework v3.1
xAIShadow mapping xAI Frontier AI Framework, 30 Jun 2026	“"xAI applies a systemic risk acceptance criteria to each identified risk, incorporating a margin of security, to determine whether each identified systemic risk and the overall systemic risk are acceptable" (s.2.3).” 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.	exactEQ confidence: high	xAI Frontier AI Framework, 30 Jun 2026
MetaShadow mapping Meta Advanced AI Scaling Framework v2	“"Residual risk: describes the level of risk that a Frontier AI model presents after mitigations have been implemented." (Appendix I)”	exactEQ confidence: high	Meta Advanced AI Scaling Framework v2
EUShadow mapping EU GPAI Code of Practice, Safety and Security Chapter	“Commitment 4, in full: "Signatories commit to specifying systemic risk acceptance criteria and determining whether the systemic risks stemming from the model are acceptable ... Signatories commit to deciding whether or not to proceed ... based on the systemic risk acceptance determination." The determination "will incorporate a safety margin" (Measure 4.1). Described in the source as the fullest, most procedurally explicit residual-risk-and-acceptance text in the corpus.”	exactEQ confidence: high	EU GPAI Code of Practice, Safety and Security Chapter
California SB 53Shadow mapping California SB 53	“"(4) Reviewing assessments and adequacy of mitigations as part of the decision to deploy a frontier model or use it extensively internally." (22757.12(a))”	closeCL confidence: high	California SB 53
Frontier Model ForumShadow mapping Frontier Model Forum: Risk Taxonomy and Thresholds / Third-Party Assessments	“"Acceptable Development or Deployment Thresholds (also called 'outcome-focused thresholds' or 'residual risk thresholds')" (s3.1). Third-party assessments: developers "evaluate their effectiveness and whether the residual risk is acceptable for proceeding with further training or deployment" (1.2).”	exactEQ confidence: high	Frontier Model Forum: Risk Taxonomy and Thresholds
AI Seoul Summit (Frontier AI Safety Commitments)Shadow mapping Frontier AI Safety Commitments (AI Seoul Summit 2024)	“Commitment IV: "only if they assess that residual risks would stay below the thresholds".”	closeCL confidence: high	Frontier AI Safety Commitments (AI Seoul Summit 2024)
AmazonShadow mapping Amazon Frontier Model Safety Framework	“"If pre-deployment evaluations demonstrate that a model has capabilities that meet or exceed a Critical Capability Threshold, the model will not be publicly deployed without appropriate risk mitigation measures" (Overview, p.1); if safeguards cannot adequately mitigate the risk, "we will not deploy the model until we have identified and implemented appropriate additional safeguards" (s.3, p.3).”	exactEQ confidence: high	Amazon Frontier Model Safety Framework

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