The EZDRM approach to Revenue Security on streaming video can be visualized as a multi-dimensional set of solutions. A growing dimension of that approach is the secure addition of C2PA provenance metadata to streaming content, offering authenticity and integrity as a standard component of added value. But that provenance data means little unless consumer devices can visualize that it exists, confirm its signature is valid, and surface the source information behind it. As a committed partner of EZDRM, VisualOn closes that loop across its family of player technologies — and this post, the latest in the Revenue Security Prism series, tells that story from both sides of the integration.
In today’s streaming landscape, the hard truth is that AI-generated and AI-manipulated video is no longer a novelty — it is mainstream, and it is putting content authenticity at the center of the streaming industry’s next big challenge. AI-generated video is projected to account for 10% of all digital video content in 2026, and by most accounts, that content is now effectively indistinguishable from authentic footage under unaided human review and by most automated detectors.[1] That erosion of trust is measurable at the individual level, too: in controlled studies, viewers correctly identify high-quality deepfake video only around 24.5% of the time — meaning audiences are fooled far more often than not, even when they know a fake may be present.[2] Regulators are responding to that gap directly: under the EU AI Act’s Article 50, deployers must disclose AI-generated content in a machine-readable format starting August 2026, with fines of up to €15 million or 3% of global turnover for non-compliance — a clear signal that content authenticity and provenance are moving from best practice to legal requirement.[3] For streaming platforms, the risk isn’t hypothetical: manipulated clips, synthetic footage, and falsely-attributed content can spread through a platform as easily as authentic programming, undermining viewer trust and brand value alike.
Detection alone can’t close this gap. Deepfake detection tools that perform well in controlled lab benchmarks routinely lose much of their accuracy in real-world deployment, and the underlying arms race — where generation models are trained against the very detectors built to catch them — means no static detection tool stays effective for long. Consequently, operators and platforms need a more proactive, multi-faceted approach: one that doesn’t just try to catch fakes after the fact, but discloses, at the point of playback, whether content has been AI-generated or modified, and traces it back to its source. How can platforms surface that information and visually assure viewers of what they’re actually watching? The answer lies in secure provenance.
The innovation: EZDRM, a global leader in content security, is offering a standards-based solution that adds signed standard C2PA provenance data to any video stream as a complement to their acclaimed DRM as a Service offering.
The method: EZDRM offers its Provenance Signature Service to enable publishers to add C2PA provenance to live and on-demand video streams. This gives media companies and service providers a standards-based method to assert origin, integrity, and chain of custody — helping to differentiate their streams as high-value trusted content. The Provenance Signature Service adds secure, C2PA validated origin metadata as an easy to integrate real-time function within existing production workflows.
To unlock its full potential to support content integrity, EZDRM’s breakthrough solution deserves seamless, native support within the multimedia player.
As experts in multimedia playback, VisualOn introduces the Open+ player suite — ExoPlayer+, AVPlayer+, and Shaka+ — engineered to be 100% API-compatible with their open-source counterparts. That compatibility is the crucial detail: provenance support doesn’t require a new SDK, a new integration pattern, or a parallel playback path. It’s the same player, same APIs, same lifecycle your app already depends on, with C2PA handling built into the pipeline underneath.
On the playback side, Open+ handles extracting and verification of the C2PA data natively. As the player ingests a stream, it parses the C2PA information from the video payload, extracts the provenance claims and signature — all during normal segment retrieval, so provenance verification rides alongside playback rather than blocking it. This works across both live and VOD delivery, and across the codec and DRM configurations Open+ already supports, so operators aren’t choosing between provenance and their existing encryption or codec strategy.
C2PA metadata validated within the VisualOn Open+ Player suite, means that the trust chain, key handling, and signature format Open+ expects are already aligned with what EZDRM’s Provenance Signature Service produces on the encoding/packaging side. There’s no custom glue code required to bridge the two systems — verification logic is built into the player’s core, not layered on as a plugin or wrapper.
For developers, there’s no massive infrastructure overhaul needed. Just drop ExoPlayer+, AVPlayer+, or Shaka+ into your existing project in place of the open-source player you’re already using, and the app inherits Open+’s extraction and sending logic automatically, with no changes to your existing playback integration, ad insertion, or DRM configuration.
For operators, this enables them to benefit from a pre-validated integration of both workflow and consumer-side components, enabling both authenticity and revenue security with minimal development costs, and from EZDRM and VisualOn joint expert support.
[1] AI-generated video projected to reach 10% of all digital video content in 2026, and widely described as indistinguishable from authentic content under unaided human review and by most automated detectors: Kapwing, 55 AI-Generated Video Statistics: Disclosure, Detection, and Trust (2026), citing Genra AI / Runway Turing Reel, SQ Magazine, and University of Florida research.
[2] Human accuracy identifying high-quality deepfake video at approximately 24.5%: Korshunov and Marcel, Idiap Research Institute, as cited in multiple 2026 industry reports including Eftsure, Deepfake Statistics 2026: Key Facts for CFOs, and StingRAI, Deepfake Statistics 2026: 40+ Verified Numbers.
[3] EU AI Act Article 50 mandatory disclosure of AI-generated content in machine-readable format, effective August 2, 2026, with fines up to €15 million or 3% of global turnover: Netarx, Deepfake Statistics 2026: AI Fraud by the Numbers; Truthscan, Deepfake Statistics 2026: AI Fraud Data and Trends.