Three separate initiatives are beginning to tackle different parts of the problem around access, attribution, and measurement of content — three things that are vital to get right before we start having serious conversations about scalable content licensing for AI. 

• CoMP, which creates the protocol for AI content transactions and licensing.

• RSL, which creates a common rights and royalty framework so AI companies can license content and compensate rights holders at scale.

• SPUR Telemetry, which creates a common way to measure how publisher content is used inside AI systems.

Taken together, they start to look less like individual projects and more like the foundations of an emerging AI content infrastructure. Let’s run through these one by one.

1. CoMP: building the content marketplace

The IAB Tech Lab’s Content Marketplace Protocol (CoMP) is focused on the transaction itself. It asks questions such as:

• How does an AI company discover content?

• How does it understand usage rights?

• How does it request access?

• How does a publisher express licensing terms?

• How do content providers and AI systems exchange information?

In simple terms, CoMP is trying to create infrastructure for a functioning content marketplace between publishers and AI companies.

If publishers want to license content at scale, there needs to be a common way to communicate rights, permissions, pricing, and access. CoMP is one of the most serious attempts to create that framework. More about CoMP here.

2. RSL: creating the licensing network

While CoMP defines how publishers and AI companies can transact, RSL is trying to create the ecosystem that makes those transactions practical.

The Really Simple Licensing (RSL) initiative provides a common framework for publishers to declare rights, participate in licensing programmes, and receive payment when their content is used. Rather than every AI company negotiating separately with every publisher, RSL aims to simplify rights management and create a scalable mechanism for collecting licence fees and distributing royalties to rights holders.

In many ways, CoMP and RSL are complementary. CoMP provides the technical plumbing for transactions; RSL provides the commercial framework that allows those transactions to happen at scale. More about RSL here.

3. SPUR: understanding what happens after the transaction

Today publishers can measure referrals, clicks, and pageviews. They cannot easily see whether an AI system retrieved their content, used it to generate an answer, cited it, or whether the user took action afterwards.

This is the problem SPUR Telemetry is trying to solve. It proposes a standard event stream for AI systems — think of it as Google Analytics for AI.

Instead of measuring Pageview → Click → Conversion, SPUR introduces a new set of signals:

• Retrieved: The AI system accessed the content.

• Grounded: The content influenced the model’s response, even if it wasn’t cited.

• Cited: The publisher was explicitly referenced.

• Engaged:The user took a meaningful action, such as clicking a citation or asking the agent to continue.

Of these, grounding is arguably the most important. It measures influence rather than attribution, helping publishers understand when their journalism shaped an AI-generated answer without being explicitly referenced. SPUR also measures activity across an entire session, providing a richer picture of how content contributes throughout a conversation. More about SPUR here.

What has not yet been solved

One very important piece of the puzzle that has not yet been solved is audience intelligence. The standards above are primarily focused on content usage, not user data. Even with the above, news organisations still lack many of the audience signals that have traditionally powered subscriptions, advertising, and audience development.

From a publisher perspective, an ideal future dataset might include information such as country, language, device type, user segment, subscriber status, professional role, interests, and repeat usage patterns. That would provide valuable context around who is consuming content through AI systems and how those interactions translate into business value.

This is also where privacy concerns quickly emerge. Aggregate signals such as country, language, and device type are relatively uncontroversial and could provide useful context while preserving user privacy. Once discussions move toward persistent identifiers, subscriber information, detailed demographics, or user-level behavioural data, the conversation becomes significantly more complex.

Why this matters now

What is encouraging is not any individual initiative but the fact that multiple efforts are emerging at roughly the same time.

• CoMP addresses content transactions and exchange.

• RSL addresses rights management and royalty distribution.

• SPUR addresses content usage and measurement.

The industry should resist the temptation to create additional/competing standards for each of these problems. Publishers, technology companies, and standards bodies will all benefit if these efforts align around common approaches and common definitions. A fragmented standards landscape will delay, or worse, any meaningful adoption from the major AI platforms.

The SPUR specification is currently open for public comment, which means now is the time for publishers to engage.

The most important standards in digital media were shaped before they became standards. This is one of those moments. If publishers want a future where content usage inside AI systems is measurable, transparent, and comparable across platforms, now is the time to help define what that future looks like.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

Banner image: Adobe stock image by Tadamichi.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .