DAM Cowboy
Search

W32 · August 2026

Infrastructure for generated content starts to look like infrastructure for trusted content

August brought a cluster of releases that share a pattern: vendors are building plumbing for AI-created and AI-moderated media at the same speed standards bodies are building plumbing for provenance and trust. The gap between generation and verification is narrowing, but it is not yet closed.

topic/video-moderationCloudinary Blog

Cloudinary extends automated moderation from images to video ↗

When we launched Cloudinary Moderation, we set out to turn static brand guidelines into automated standards that every asset entering your pipeline has to meet. Until now, that meant images. Today, we’re extending moderation to the format that’s growing fastest and is hardest to review manually: video. Video moderation is now generally available in Cloudinary. […] The post Video Moderation is Here: How Cloudinary Reviews Every Video Against Your Standards appeared first on Cloudinary Blog.

Read on Cloudinary Blog →

TL;DR · Why it matters

Brand guidelines that used to require manual review now run automatically against every uploaded video. The feature scans for content violations, format compliance, and brand-standard adherence frame by frame. Video is the format growing fastest and hardest to moderate at scale, so workflow automation here matters more than in static formats. This is less about catching bad content and more about enforcing consistency before assets enter distribution.
Permalink & share
Bynder Blog
topic/facial-recognition-searchBynder Blog

Facial recognition search in DAM: how it works and why it matters ↗

TL;DR · Why it matters

Bynder explains the mechanics of facial recognition in asset libraries: biometric mapping, similarity scoring, and privacy boundaries. The use case is narrow (find all images of a specific person across thousands of untagged files) but the accuracy threshold is high. This only works when the DAM holds enough samples per face to train the model and when metadata practices are clean enough to validate results. Facial recognition does not replace tagging; it accelerates tagging for one specific dimension.
Permalink & share
topic/image-generation-apiCloudinary Blog

Cloudinary ships multi-model image generation API with built-in asset management ↗

At conference booths, developers often ask whether we support image generation at Cloudinary, given our emphasis on media management. As of now, I can say, “Yes! Yes, we do.” — here’s how! Cloudinary’s Image Generation API lets developers generate images from text prompts using multiple AI model families, then store the result as a managed […] The post Cloudinary’s New Image Generation API: One API, Multiple AI Models, and Built-in Asset Management appeared first on Cloudinary Blog.

Read on Cloudinary Blog →

TL;DR · Why it matters

The new API lets developers generate images from text prompts using multiple AI model families, then store and manage those images inside Cloudinary's existing DAM infrastructure. This is not a standalone generation tool; it is generation wired directly into the asset pipeline. The architecture assumes generated images will be versioned, tagged, and distributed like any other asset, which is the correct assumption. Generated content that bypasses the DAM creates a shadow library that no one can govern six months later.
Permalink & share
topic/media-provenance-summitIPTC

IPTC releases video overview of Toronto media provenance summit ↗

The IPTC is pleased to share with the public a video overview of the IPTC Media Provenance Summit which was held on Thursday 16th April in Toronto, Canada. The video shows interviews with speakers from the event, including Angela Pacienza of the Globe & Mail, Dan Dzuban of Sony and the C2PA Steering Committee, Judy […] The post Video overview of Toronto IPTC Media Provenance Summit appeared first on IPTC.

Read on IPTC →

TL;DR · Why it matters

The summit covered content authenticity work from news organizations, camera manufacturers, and the C2PA steering committee. Interviews with speakers from the Globe & Mail, Sony, and the C2PA show convergence around a shared goal: baking provenance into the capture and distribution chain, not applying it retroactively. This is infrastructure work, not feature work. The video is worth watching if you manage media workflows that will need to prove authenticity under regulatory or editorial pressure in the next 24 months.
Permalink & share
topic/ai-in-damAcquia Blog

What AI can (and cannot) do for your DAM ↗

Auto-tagging, facial recognition, video transcription — AI DAM features only work when your foundation is clean. Learn the three kinds of AI in your DAM world and how to use them well.

Read on Acquia Blog →

TL;DR · Why it matters

Acquia maps three categories of AI in the DAM world: assistive (auto-tagging, transcription), augmentative (search expansion, similarity), and generative (content creation). The piece is correct that all three categories depend on clean foundational data. Auto-tagging a messy library produces messy tags at scale. The best use of AI in DAM is not to fix broken workflows; it is to accelerate workflows that already work. If your taxonomy is inconsistent, AI will amplify the inconsistency.
Permalink & share
topic/newscodes-updateIPTC

IPTC releases Q2 2026 NewsCodes vocabulary update ↗

The NewsCodes Working Group is pleased to release the Q2 2026 update to our NewsCodes vocabularies. The first update since October 2025, this update is small because the NewsCodes Working Group is working on a large update to the sports branch of Media Topics that is not yet complete. Changes in this release: New Media […] The post IPTC NewsCodes 2026-Q2 update released appeared first on IPTC.

Read on IPTC →

TL;DR · Why it matters

The update is small because the working group is preparing a larger sports branch revision for Media Topics. NewsCodes are controlled vocabularies for news metadata, used by publishers to tag, classify, and route content. If your DAM ingests wire service content or publishes to syndication partners, your metadata schema likely maps to NewsCodes somewhere in the pipeline. The Q2 update maintains compatibility while the sports taxonomy gets rebuilt. Watch for the sports release in Q3 if you manage sports media or event-based content at volume.
Permalink & share

Resource of the week

How to tell if your DAM is actually working

Six diagnostic questions that separate functional DAMs from expensive file cabinets. Run this before your next quarterly review.

Run the diagnostic →

Field Notes

  • StandardsIPTC NewsCodes Q2 2026 update released (small; larger sports taxonomy revision in progress)
  • ConferenceIPTC Media Provenance Summit video now public (Toronto, April 2026)

What are you building to bridge creation and trust in your own pipeline?