For publishers, getting cited in AI chatbot responses isn’t enough. They also want those answers to reflect their editorial judgement and the culture of the communities they cover. A new initiative called SAIL – which stands for Standardized Agentic Intelligence Ledger – aims to do both, compensating publishers when AI scrapes their content and guaranteeing the outputs adhere to the same cultural standards they apply to their own coverage.
The framework was designed by AI licensing platform Next Net in partnership with Sundial Media & Technology Group, the publisher of Essence, Refinery29 and Afropunk, among others.
SAIL is a digital record-keeping system that tracks how AI solutions use publisher content and mix it with other sources, said Sundial CEO Kirk McDonald. It’s meant to protect the value of high-quality publisher content when it’s cited by AI alongside less rigorous – but still culturally relevant – user-generated content.
Think of the framework as a more collaborative alternative to striking one-off licensing deals with AI vendors, McDonald said, or to the nuclear option of suing them over unauthorized content scraping.
Tech and human connections
Publishing brands have built authentic connections with their audiences over decades, and they deserve a way to maintain them in the AI era.
Sundial’s brands, for example, are respected as some of the top resources for news about the Black community, McDonald said. When it comes to Black women in particular, he added, Essence is often cited in AI responses as the most authoritative media source for topics relevant to that audience. So Sundial wants to ensure that authority is protected – and appropriately valued by AI partners – as AI usage gains momentum.
The risk is that AI can misinterpret or flatten content, McDonald said, especially when professionally-produced work is commingled with user-generated content that may not be as culturally sensitive and may perpetuate harmful stereotypes.
To address that risk, SAIL provides a technical bridge between AI vendors and a publisher’s own internal AI tools that have already been trained with cultural sensitivity in mind.
For example, Sundial has developed its own marketing automation and research intelligence platform, a content creator platform and a consumer-facing chatbot interface. In training those tools, McDonald said, Sundial ingested a century’s worth of combined cultural expertise from across its publishing brands. By integrating internal systems with third-party AI tools, SAIL bakes Sundial’s cultural understanding directly into third-party training models.
Because Next Net’s platform runs on the same NVIDIA infrastructure used by many third-party AI vendors, it can more easily facilitate those connections, said Next Net CEO Franklin Rios. Since January, Next Net has been “vectorizing” Sundial’s content catalog, he said, which involves translating it into software language that AI models can understand.
But the tech piece is just one part of SAIL. The framework is also meant to give publishers a seat at the table when it comes to how AI tools are trained, McDonald said.
“If I can’t be involved in the design of the outcome,” he said, “then I’m just a supplier of content and the machines are leading us.”
Giving publishers their due
But how does the framework actually generate revenue for publishers?
SAIL tracks each time an AI system draws on a publisher’s content to formulate an answer, and the SAIL ledger is compatible with the IAB Tech Lab’s CoMP AI framework, which uses a “pay-per-query” model to pay publishers whenever their content is used in a response.
Next Net gets a cut of any revenue returned to the publisher, with the specific terms negotiated on a case-by-case basis.
But SAIL is about more than just paying publishers for their content, McDonald emphasized. They’re also being compensated for the time and effort they’ve invested to become authoritative voices in their communities, he said, which is the kind of work automated systems often overlook.
Because the framework tells AI systems to treat certain publishers as more authoritative sources, those outlets are more likely to be cited in AI responses which, in turn, drives referral traffic to the original source if users click on the citations.
Collaboration over conflict
But referral traffic, while nice, isn’t as important as having a say in how AI-driven content discovery works, McDonald said.
What publishers really need, he said, is a “source file” that shows how an AI answer was assembled, including which resources it pulled from the publisher – an article, video and insights from a live event, for example – as well as which other sources helped shape the AI’s response.
This accountability gives publishers the power to put their seal of approval on AI outputs. “I don’t think anyone wants their work embellished or compromised in a way that they wouldn’t endorse,” McDonald said.
This model is also a welcome alternative to the increasingly adversarial dynamic between media owners and AI vendors, added McDonald, who has experience on both sides of the marketplace, including as the former CEO of GroupM North America, chief business officer of Xandr during the AT&T days and as president of PubMatic, where he worked with publishers.
Having seen firsthand how automation and competing interests can leave different stakeholders underpaid for the value they create, he believes the best way for publishers to protect the authority and value of their content is to work collaboratively with AI companies, rather than fight them.
“You start doing these David versus Goliath arguments – I don’t actually want to slay the giant,” McDonald said. “What I would like us to do is come as collaborators to the table, design an experience that’s tackling any unintended bias and introduce technology that holds us all accountable.”