3AUG

SMART Stories shows its CRDT working

4 min read

17:21UTC

An EVS-led whiteboard at mpts.london on Thursday 14 May revealed the SMART Stories Story Object Model as a conflict-free replicated data type, six weeks after the consortium launch on 30 April, with Trint contributing a live-news interface mock and a full proof of concept targeted at IBC 2026.

← Back to AI marking code gets 190 names, no broadcasters Jump to analysis ↓

IndustryDeveloping

Key takeaway

SMART Stories revealed a CRDT-based Story Object Model at mpts.london on 14 May, with IBC 2026 the proof-of-concept deadline.

SMART STORIES showed its first public architectural detail at the mpts.london conference on Thursday 14 May, six weeks after the consortium launch announcement on 30 April . An EVS-led whiteboard, photographed during the talk, sets out the Story Object Model (SOM) as a conflict-free replicated data type (CRDT), the same data structure that powers Figma and Google Docs, synchronising news context across Local and Regional altitudes within a newsroom and between newsrooms.

The runtime architecture splits into three regions. LIVE ingests feeds through a router into Ksender, Kvideo and IVR modules, with a Cerebrum reasoning agent and ViaMap and Social sub-skills sitting above the router. NEARLIVE is the store layer, with Moments Lab indexing records at two altitudes: a Local SOM per newsroom and a Regional SOM per region, sync arrows pointing both ways. DISTRO runs Skills against the store and feeds Web, Bing and Trace output adapters.

Trint contributed a live-news interface mock, framed as "Breaking News, Croatia Interview", that shows four SOM behaviours: known story state, AI-monitored transcript matching against the current premise, context-based flagging of new facts (the example used was "death toll now 12") with verification-required tagging, and a Live Quote Finder that surfaces soundbites against organisation-specific Skills plus the current story context.

SMART STORIES is selling Skills as the moat. The pitch from the stage was: "BBC Skills run BBC standards. ITN Skills run ITN's. Framework open. Workflow yours." Read as competitive strategy, this puts an open protocol layer underneath proprietary editorial DNA, the inverse of the Avid plus Google Cloud stack announced at NAB 2026 , which buries a single vendor's intelligence layer underneath proprietary newsroom output. Skills carry the moat: shared transport, defended editorial judgment, which means broadcaster IP stays inside the building while transport interoperates. AP, juggling concurrent newsroom buyouts , is a full consortium member; the editorial test is whether the open layer survives contact with members whose first-party labour position is shrinking.

Governance and transport remain open questions. Trint's framing of an editor "publishing back to the bus" implies a pub/sub message layer, which puts the choice between MQTT, NATS or a custom protocol on the standards committee, and the choice between SMPTE, AMWA NMOS, EBU or a new body on the governance question. The full proof of concept is due at IBC 2026 in Amsterdam, 11 to 14 September.

Deep Analysis

In plain English

SMART Stories is a project involving nine major news organisations, including the BBC, AP, NBCUniversal and Sky, to build a shared technology standard for AI-assisted news production. Think of it as agreeing on a common file format for news stories so that AI tools from different companies can all read and update the same story without conflicts. On 14 May they explained in public for the first time exactly how that shared format works technically. It uses a data structure called a CRDT, the same technology that powers Google Docs' real-time collaboration, so that multiple AI agents can edit the same story simultaneously without overwriting each other's work.

Deep Analysis

Root Causes

The CRDT architecture reflects a core structural problem in multi-organisation agentic production: traditional databases require a single source of truth, which in a consortium setting implies a single controlling entity.

The SMART Stories consortium spans BBC, ITN, AP, NBCUniversal, Channel 4, Sky, ITV, Al Jazeera and The Washington Post: all compete editorially and cannot cede data sovereignty to a single vendor or member. A CRDT-based SOM solves this by making the story object self-resolving: any node can write, and the CRDT merge function determines the canonical state without requiring a central arbiter.

The three-layer runtime architecture (LIVE / NEARLIVE / DISTRO) is also a response to a structural newsroom reality: different production speeds require different data freshness guarantees. LIVE needs sub-second synchronisation for breaking news; NEARLIVE can tolerate seconds to minutes; DISTRO (Skills feeding web and syndication outputs) can tolerate minutes to hours. A single data layer cannot serve all three simultaneously without throughput tradeoffs.

What could happen next?

Meaning
The full proof of concept is due at IBC Amsterdam on 11-14 September 2026. If the consortium demonstrates a working multi-organisation CRDT at IBC, the pressure on Avid and Adobe to publish NMOS-compatible APIs for their agentic editing products will intensify materially before year-end. AP's concurrent buyout of journalists (ID:3190) is the stress test: whether an open-standard agentic layer can absorb the editorial workload of a reduced human newsroom is the question the IBC demo must answer.

Source Landscape

This story draws on centre-left-leaning sources from United States

United States

Primary parallel: The MOS protocol (Media Object Server, 1998) established the direct precedent. MOS was an open standard developed by a consortium of nine newsroom technology vendors (including ENPS and iNews) to enable news rundown data to move between editorial systems and playout servers without point-to-point integrations.

It became the broadcast interoperability standard within five years of launch and is still in use in the majority of broadcast newsrooms globally. SMART STORIES is attempting the same structural move: an open-standard data model for story objects that sits above any individual vendor's editorial system.

Counter-parallel: The DVB consortium, formed in 1993 to standardise digital video broadcast transmission across European broadcasters, succeeded at the transport layer but failed to produce a unified application-layer standard for interactive services (DVB-MHP).

Interactive TV standards fragmented into national implementations because editorial workflows vary too much between organisations for a single application standard to hold. SMART STORIES' Skills architecture, which deliberately makes editorial logic per-organisation, appears designed to avoid this failure mode.

SMART STORIES is an IBC Accelerator project with seed funding, not a commercial product. The economic analysis sits at the vendor level: the nine consortium member organisations collectively purchase hundreds of millions of pounds in news production technology annually, and the open-standard architecture they are building represents a direct threat to the closed-vendor revenue model of Avid and Adobe .

If the SOM becomes an adopted standard, the value capture moves from the data-layer vendors (who currently charge for proprietary story-management integrations) to the Skills layer (per-organisation editorial logic, which each member builds internally) and the application-layer vendors (Trint for transcription, Cerebrum for reasoning, ViaMap for geospatial). That revenue redistribution is the commercial stake behind the architecture choice.

Consensus view: SMPTE's open-standards working groups and the AMWA NMOS (Networked Media Open Specifications) community have both identified the central challenge in agentic news production interoperability as state management rather than transport: multiple AI agents editing and annotating the same story object simultaneously need a conflict-resolution mechanism.

The SMART STORIES team's choice of a CRDT (conflict-free replicated data type) is the architecturally correct answer, directly adapted from the distributed-systems patterns that power Figma's real-time collaborative editing and Liveblocks' collaborative document infrastructure.

Counter-view: IBC technical programme contributors from Avid's engineering team have noted in prior years that CRDT-based synchronisation degrades under high-frequency concurrent writes, which is exactly the condition a breaking-news workflow generates: dozens of agents tagging, annotating and updating the same story object within seconds of a developing event.

Figma's CRDT implementation caps real-time collaborators at around 100 for this reason; a major newsroom's agentic production floor could generate an order of magnitude more concurrent writes.

Key tension: Whether the SMART STORIES CRDT-based SOM can sustain conflict-free synchronisation at the write volume and latency requirements of live breaking-news workflows, or whether the architecture will require a centralised coordination layer that reintroduces vendor dependency.

Sources:Bloomberg

Mentions:SMART STORIES →EVS →Story Object Model →Trint →Moments Lab →mpts.london →Associated Press →British Broadcasting Corporation →ITN →IBC 2026 Accelerator →Amsterdam →Google →Google Cloud →NBCUniversal →SMPTE →Avid →Croatia →SOC →

First Reported In

Update #2 · BuzzFeed sold for AI; agentic ads go primetime

Bloomberg· 17 May 2026

Read original →

Causes and effects

Caused by

Nine newsrooms back SMART STORIES open standard

mpts.london architectural reveal advances the SMART Stories consortium launch of 30 April with first public SOM CRDT specification detail.

Occurred 30 Apr 2026

Read story →

AP runs SMART STORIES and journalist buyouts in same fortnight

AP's dual role as consortium member and concurrent buyout subject is the editorial stress test for SMART Stories open-layer openness.

Occurred 15 Apr 2026

Read story →

Avid embeds Google Gemini in Media Composer

SMART Stories open-protocol Skills layer is the inverse of the Avid/Google Cloud closed-vendor stack: shared transport, defended editorial judgment.

Occurred 16 Apr 2026

Read story →

This Event

SMART Stories shows its CRDT working

Pins the architectural bet: open protocol underneath proprietary editorial DNA, the inverse of the Avid-Google Cloud closed-vendor stack and the test of whether shared transport survives contact with consortium members under buyout pressure.

Different Perspectives

Italian market: Fastweb signed, Mediaset and RAI did not

Fastweb, the Italian broadband operator, signed Section 2 as a deployer, while Mediaset and the state broadcaster RAI are both absent. Italy therefore shows the split in miniature: a telecommunications company with one obvious deployment surface accepted the standard, and the two companies that actually broadcast to Italian audiences did not.

Paramount Skydance and its French exposure

Paramount filed a stipulation not to close on 24 July and had trial specialists admitted three days later, after nine months arguing that antitrust risk was regulatory rather than judicial. Its 22 July European clearance obliges it to divest a stake in the Universal International Pictures distribution venture and to strike no Universal distribution deal in Europe for a decade.

Time and the agent-advertising camp

Time and Mobian treat AI agent traffic as sellable inventory rather than leakage, on a site where bots now outnumber humans most days. Chief operating officer Mark Howard frames agent impressions as an extension of sponsorship, against an industry consensus still pursuing compensation for scraping.

German broadcasters

ARD, ZDF, RTL and ProSiebenSat.1 are all absent from the list, making Germany the largest single national bloc of non-signatories. Their position rests on three defensible grounds: signature is voluntary, most deployed systems are grandfathered to 2 December 2026, and a broadcaster's compliance surface spans production, archive, advertising and distribution rather than one product. None has publicly explained the decision.

Synthesia and the signed vendor layer

Synthesia signed Section 1 alongside Getty Images, Google, Meta, Microsoft, Anthropic, OpenAI, Mistral and Aleph Alpha, accepting a documented marking standard for the product they sell. London-headquartered Synthesia is the only pure AI-video vendor on the providers list, and gains a compliance artefact to put in front of European media buyers at the moment those buyers acquire a live obligation.

European Commission

The Commission published its initial signatory list on 31 July with roughly 190 organisations and let Article 50 take effect on 2 August as scheduled. Its own page stresses that adherence to the Code is voluntary while the transparency requirements are legal obligations, which frames non-signature as an evidentiary choice rather than a breach.