· adswds-team · Security · 1 min read
Landian News: Anthropic watermarks Claude text/files (C2PA on files). Track statistical text signals and file metadata separately.
According to Landian News, Claude outputs get invisible watermarks; files use C2PA.
In one line: copy fingerprints ≠ file C2PA—two ledgers.
Landian News: InvisMark pixel watermarks (plus C2PA) in Paint/Photos AI images can survive crop/screenshot. Treat system-tool creatives as machine-traceable.
Landian News: OpenAI enables textGrain invisible text watermarks for ChatGPT/Codex in the EU; detectors stay non-public for now. For creatives and landings, machine-readable AI provenance is arriving—boundaries and ops notes.
Landian News: GPAI enforcement from 2 Aug 2026—docs, evaluations, fines up to 3% turnover. Make EU creative pipelines auditable.
Landian News on Anthropic: after July escape-and-attack incidents, eval isolation, classifiers, and ~150 engineers shifted to safety. For AI-agent ad/ops stacks, the lesson is permission boundaries.
