extremism
Enable an AI agent to assess a claim of extremism against an explicit definition, distinguish beliefs from conduct and threats, and identify evidence-supported, proportionate next steps.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
Researched by: Codex
Purpose and description
Enable an AI agent to assess a claim of extremism against an explicit definition, distinguish beliefs from conduct and threats, and identify evidence-supported, proportionate next steps.
It can be Assess an extremism attribution under a named framework and explain which criteria are supported, contradicted or unresolved.; Compare assessments under different frameworks without silently merging their definitions.; Trace an allegation to contextualized evidence and distinguish endorsement from quotation, association or hostile attribution.; Separate ideological classification from a conduct or threat assessment and route each to the appropriate process.; Identify proportionate next steps within an agent's authority, including further inquiry, correction, review or referral based on independently evidenced harm.; Reassess, narrow or withdraw an attribution when evidence, conduct or the applicable framework changes..
Distinguishing features
A classification must satisfy stated extremism criteria; distance from majority opinion, novelty or unpopularity alone does not establish it.
A proposal for fundamental change must be assessed separately from its stance toward coercion, violence, opponents and equal standing; radical ambition alone does not settle the classification.
Quoted, reported, satirical or critically examined extremist material must be distinguished from endorsement through evidence of communicative context.
Belief, advocacy, preparation and action must remain distinguishable; evidence for one does not automatically establish the others.
An extremism assessment and a terrorism, violence or illegality assessment must each meet their own criteria; neither classification automatically supplies the other.
Scope
+ The definition, reference context and threshold used to classify something as extremist
+ Attributed commitments concerning opponents, pluralism, equal standing and permissible means
+ Evidence distinguishing expressed beliefs, endorsed actions, preparation and observed conduct
+ The reach, persistence and variation of extremist commitments within an assessed subject
+ Changes in the assessment and the conditions for proportionate intervention or reassessment
- Complete biographies, identities and organizational structures of assessed actors
- General models of political, religious or philosophical ideologies
- Full incident records and operational investigation of violence or terrorism
- Criminal liability, proscription and jurisdiction-specific legal adjudication
- General models of protest, civil disobedience, polarization and radical social change
- Platform-wide content moderation policies and enforcement workflows
Characteristics
- Assessment subject
- statement | position | practice | individual | group | network | institution Sets the unit to which evidence and conclusions apply and prevents a statement-level judgment from becoming an unsupported judgment about an entire population.
- Classification framework
- Named definition, issuing authority or author, version, criteria and applicable context Makes the meaning of extremism inspectable instead of treating the label as self-explanatory.
- Classification status
- unassessed | insufficient evidence | criteria met | criteria not met | contested; relative to a specified framework Separates unresolved evidence from a supported positive or negative assessment.
- Normative reference
- Explicit reference to political norms, legal standards, rights principles or another stated baseline Exposes which baseline makes the position extreme and allows that baseline itself to be challenged.
- Stance toward opponents
- Record separately: recognition of equal standing, acceptance of opposition, exclusion, subordination, elimination, mixed or unknown Captures the concrete treatment envisaged for disagreement without reducing all ideological differences to extremism.
- Relationship to coercion and violence
- Separately evidenced rejection, conditional acceptance, endorsement, threat, preparation or enactment; mixed or unknown permitted Supports distinctions between classification, harmful conduct and a credible threat.
- Evidence coverage
- Number of independently corroborated observations, observation period and proportion of required criteria supported Shows the extent of the evidence while avoiding the assumption that repeated copies of one allegation are independent corroboration.
- Attribution reach
- individual expression | factional position | leadership position | formal policy | demonstrated collective practice | disputed | unknown Prevents unsupported generalization from members to organizations or from organizations to associated communities.
- Observed trajectory
- stable | intensifying | moderating | internally diverging | disengaged from specified conduct | unclear; over a stated interval Enables time-bounded reassessment without assuming a fixed identity or inevitable progression toward violence.
Also called
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 5 bundles · 10 layers · 10 findings · 20 questions.
Classification and boundaries Establishes what extremism means in this assessment and what would count against applying the label.
Without an explicit framework and exclusions, the label can collapse disagreement, radical change and harmful conduct into one judgment.
Definition and reference
Locates the definition, its authority and its normative reference.
Applicable extremism criteria
Record the chosen definition, its necessary or sufficient criteria, and why it applies to the subject and context.
- Which definition is being used, and does it turn on ideological distance, treatment of others, permissible means or a stated combination? definition
- Who supplied this definition, for what purpose, and which version and context govern its use? provenance
Exclusions and disagreement
Tests competing interpretations and boundaries with legitimate dissent and radicalism.
Boundary and counterevidence
Record why the subject does or does not cross the chosen threshold, including evidence supporting an alternative interpretation.
- What evidence distinguishes this case from unpopular belief, peaceful radical reform, civil disobedience or forceful criticism? boundary
- Which criterion fails or remains unresolved, and would another relevant framework classify the same evidence differently? boundary
Commitments and permissible means Examines the envisioned social order, treatment of opposition and means regarded as acceptable.
An assessment needs concrete commitments that explain the classification rather than inference from ideological labels or affiliations.
Opposition and equal standing
Records how the subject proposes to treat disagreement and people outside its preferred community.
Treatment of opponents
Record explicit positions and demonstrated practices concerning opponents' participation, rights and continued existence.
- Does the subject recognize opponents' equal standing and ability to organize, dissent and participate, and what evidence supports that reading? measurement
- Does hostile language concern criticism of ideas, exclusion from participation, subordination or elimination of people, and where is the interpretation uncertain? boundary
Justification of coercion
Distinguishes objectives from the means endorsed to achieve them.
Endorsed means and limits
Record whether coercion or violence is rejected, accepted conditionally or endorsed, including stated targets and limits.
- Which coercive or violent means are explicitly endorsed or rejected, against whom, and under what stated conditions? definition
- What establishes endorsement rather than description, fictional speech, quotation or an interpretation imposed by an opponent? provenance
Evidence and attribution Connects extremism claims to contextualized evidence at the correct level of attribution.
Misquotation, circular allegations and generalization from association can produce a classification unsupported by the subject's actual commitments or conduct.
Expression in context
Establishes authorship, meaning and evidential independence of relevant expressions.
Contextualized expression
Record the original expression, surrounding exchange, authorship confidence and any material translation or interpretation choices.
- Can the expression be traced to an authentic original with enough context to assess authorship, audience, irony and quotation? provenance
- Which observations independently corroborate the interpretation, and which merely repeat the same allegation or excerpt? measurement
Individual and collective reach
Controls movement between claims about statements, members, leaders and collectives.
Supported attribution level
Record whose position is evidenced and whether a collective adopted, tolerated, rejected or disputed it.
- What supports attribution to the individual, faction, leadership or whole organization rather than to an associated person alone? boundary
- What evidence shows formal adoption, implementation, internal dissent, repudiation or sanctions concerning the position? provenance
Conduct and change Separates attributed commitments from conduct and tracks changes without assuming an inevitable escalation sequence.
An ideological classification alone cannot establish preparation for harm, and a historical assessment may no longer describe the subject.
Belief to conduct distinction
Records supported connections and gaps between expression, mobilization and action.
Observed conduct and threat evidence
Record actual advocacy, intimidation, preparation or violence separately from beliefs and from any inferred future threat.
- Which acts are directly evidenced, and which connections between beliefs, advocacy, preparation and action remain inferred? measurement
- Is there independently supported evidence of intent, capability, a target or timing that requires a separate threat assessment? action
Persistence and revision
Tracks continuity, internal variation and evidence that changes the assessment.
Time-bounded trajectory
Record changes in commitments and conduct, distinguishing ideological moderation from withdrawal from particular activities.
- Across what period and comparable observations has the subject intensified, moderated, remained stable or diverged internally? measurement
- Does new evidence establish changed beliefs, cessation of harmful conduct, a public retraction or only a change in presentation? boundary
Response and contestability Connects the assessment to authorized action while preserving correction, challenge and review.
The extremism label must not substitute for the evidence and authority required to justify a consequential response.
Grounds for response
Identifies the specific concern, governing authority and proportionate response options.
Action-specific justification
Record why a proposed response follows from supported evidence and applicable authority, including whether uncertainty requires further inquiry.
- What specific conduct, credible threat or applicable rule justifies the proposed response beyond the extremism label itself? action
- Which response is proportionate to the evidenced concern, lies within the agent's authority and requires human review before implementation? action
Challenge and expiry
Makes the attribution revisable and limits continued reliance on stale or disproven evidence.
Review and correction conditions
Record how the subject or a reviewer can challenge the attribution and what triggers narrowing, withdrawal or renewed assessment.
- What counterevidence, contextual correction or framework challenge would narrow or overturn this assessment? boundary
- When must the assessment be reviewed, and how will a correction reach decisions or records that relied on it? action
What the second pass must settle
- Which definitions should Vercy support, and can any criteria serve across contexts without concealing substantive disagreement about extremism?
- How should the model distinguish opposition to a prevailing political order from opposition to equal standing or pluralism when the prevailing order is itself exclusionary?
- What evidence thresholds are appropriate for attributing extremism to a statement, an individual or an entire organization?
- Which indicators of changing commitments or conduct have sufficient empirical support to inform reassessment without implying inevitable progression toward violence?
- How should framework-specific legal designations, disputed translations and politically motivated allegations be represented without converting authority or repetition into evidential certainty?