Source Authority Vector

Quantifies the depth of topical knowledge demonstrated by a source. This axis evaluates domain-specific credentials, the author's publication history, and the technical accuracy of prior content.

• Credential Verification: Parses structured author markup (e.g., author.type in Schema.org) to validate professional qualifications.
• Topical Consistency: Measures the semantic distance between a source's historical content corpus and the current claim. A source writing exclusively about oncology receives a high expertise score for cancer-related claims but a low score for automotive engineering.
• Peer Citation Analysis: Evaluates how frequently the source is cited as authoritative by other high-expertise entities within the same knowledge domain.

Measures the degree to which a source presents information without commercial, ideological, or sensationalist bias. This axis is critical for AI models prioritizing balanced, factual responses.

• Sentiment Variance: Analyzes the distribution of emotionally charged language versus neutral, declarative statements. High variance often signals bias.
• Sponsorship Transparency: Detects and weights the presence of sponsor or funding markup. Undisclosed commercial relationships negatively impact the index.
• Multi-Perspective Coverage: Rewards sources that explicitly reference and fairly represent competing viewpoints or contradictory evidence on a topic.

A longitudinal metric that tracks a source's factual track record over time. This component is updated continuously as claims are verified or refuted against a ground-truth knowledge base.

• Claim-Outcome Alignment: Compares past factual claims made by the source against subsequent real-world events or scientific consensus. A source that predicted a market crash accurately receives a positive update; one that propagated a debunked health myth receives a penalty.
• Retraction Velocity: Measures the time elapsed between a factual error's publication and its formal correction or retraction. Faster corrections mitigate negative impact.
• Temporal Decay: Applies a weighted decay function so that recent accuracy is more influential than historical performance from a decade ago.

Evaluates the verifiability and machine-readability of a source's origin chain. High transparency allows AI systems to cryptographically validate content integrity and authorship.

• C2PA Compliance: Checks for the presence and validity of Content Credentials, which cryptographically bind provenance metadata to the asset at creation.
• Attribution Depth: Rewards content that links claims directly to primary sources (e.g., raw datasets, original research papers) rather than secondary or tertiary summaries.
• Provenance Graph Completeness: Assesses whether the source provides a full, auditable record of its data's journey, including all transformations and intermediary agents, ideally expressed via the W3C PROV standard.

Distinguishes between genuine authority signals and artificially inflated metrics. This axis filters out manipulative practices like link farms, bot traffic, and coordinated inauthentic behavior.

• Traffic Source Entropy: Analyzes the diversity and organic nature of referral traffic. A site receiving 90% of traffic from a single, low-quality domain is flagged.
• Citation Graph Integrity: Evaluates the authority of linking domains. A link from a verified .gov or .edu domain carries exponentially more weight than a link from a known content farm.
• Bot Activity Filtration: Uses behavioral analysis to discount engagement signals (clicks, shares, dwell time) originating from non-human or scripted agents.

Models the relevance half-life of information based on its topic domain. This axis ensures that an authoritative source from 2010 is not incorrectly ranked above a current source for a fast-moving topic.

• Domain-Specific Decay Curves: Applies different decay rates. Medical research may have a half-life of 3 years, while a breaking news story has a half-life of hours. Stable knowledge (e.g., mathematical proofs) decays negligibly.
• Last-Updated Heuristic: Prioritizes content with explicit dateModified signals, but cross-references the nature of the update. A minor typo fix does not reset the decay curve as significantly as a substantive data revision.
• Event-Based Resets: Detects when a major external event (e.g., a new law, a scientific breakthrough) fundamentally alters the factual landscape, triggering an immediate recalculation of freshness for all related sources.

An algorithmic method for assigning a quantitative score to a source-citation pair, reflecting the model's certainty that the source supports the claim. This score directly feeds into the Source Authority Vector by providing a dynamic, claim-level trust metric.

• Combines semantic similarity with historical accuracy data
• Uses entailment models to verify logical support
• Flags unsupported attributions before generation

Structured data, often embedded via the W3C PROV model, that describes the origin, authorship, and transformation history of a digital asset. This metadata provides the foundational signals—such as authorship credentials and modification timestamps—that populate the Source Authority Vector.

• Implements JSON-LD and C2PA standards
• Establishes a verifiable chain of custody
• Enables automated trust assessment by AI crawlers

Cryptographically signed digital credentials that verify a specific attribute or claim about a piece of content, such as its origin or a timestamp. These tokens serve as high-weight signals within the Source Authority Vector, providing mathematically verifiable proof of authenticity.

• Uses digital signatures to prevent forgery
• Binds organizational identity to published content
• Integrates with Content Credentials ecosystems

The computational task of resolving which specific entity a citation refers to when the name is ambiguous. Accurate disambiguation is critical for the Source Authority Vector to correctly aggregate reputation signals to the right entity rather than a namesake.

• Leverages knowledge graph IDs (e.g., Wikidata QIDs)
• Resolves author name collisions in academic literature
• Prevents authority dilution across unrelated entities

The automated monitoring process that identifies when a cited source has been updated, retracted, or altered, causing a misalignment with the original claim. This mechanism maintains the temporal integrity of the Source Authority Vector by dynamically downgrading sources that have become unreliable.

• Monitors for retractions and corrections
• Triggers re-verification workflows
• Prevents citation of superseded research

Embedding explicit markers of certainty, source quality, and data freshness within content to guide an AI model's trust assessment. These signals directly inform the objectivity and freshness dimensions of the Source Authority Vector.

• Includes confidence intervals on statistical claims
• Marks primary vs. secondary sourcing explicitly
• Uses temporal validity markers for time-sensitive data