INMA: How The Guardian is taking charge of AI licensings

The Guardian is amongst the news media companies taking an assertive approach to protecting and commercialising content in a digital landscape increasingly being shaped by artificial intelligence.

During the recent INMA Webinar, Unlocking AI content licensing, members heard from Robert Hahn, director of business affairs and licensing, who explained how The Guardian has taken a proactive deal-making position to ensure it has control over its future and how the content is used in the AI era.

As one of the most prominent English-language digital news publishers globally, The Guardian now gets about 70% of its digital traffic from outside the U.K. With an archive including more than 20 million articles spanning 200 years, millions of images, and thousands of hours of audio and video, it offers a valuable source of data for AI developers — as well as raising concerns about unauthorised usage and generating new opportunities for licensing.

The Guardian’s licencing operation spans over 100 countries and covers such key areas as AI licencing, platform licencing, databases and archives, media monitoring sales to other media companies, education, and one-off sales.

Licensing is one of The Guardian’s two core revenue streams, alongside direct consumer subscriptions and advertising.

On the B2B front, the company sells content to other platforms and publishers: “It’s similar to B2C in the commercial licencing — we are focused on building subscription relationships based on recurring revenues, and we have a number of longstanding relationships in place,” Hahn said.

He noted that the licencing operation spans over 100 countries and covers such key areas as AI licencing, platform licencing, databases and archives, media monitoring sales to other media companies, education, and one-off sales.

“We also have a content development arm that seeks to build new intellectual property based on our journalism for reformatting our journalism into books and selling our content into TV and film for adaptation,” Hahn said.

AI licensing is becoming an increasingly central component of this strategy. For the current financial year, The Guardian expects AI licensing to account for roughly 40% of its licensing revenue, reflecting both the urgency and opportunity of this rapidly evolving domain.

Looking for a strategic fit

To ensure a consistent approach, Hahn said The Guardian has established clear guidelines that it uses to evaluate each licencing opportunity. The licencing must align with the company’s broader business objectives, and each deal faces scrutinisation through five organisational tests: referral traffic, customer relationship, reader revenues, other revenue types (such as licensing), and reputational value.

“A platform can fail one or two of these tests, but the balance of the benefit has to be in The Guardian’s favour as measured against these tests,” Hahn explained.

As referral traffic from platforms becomes less dependable, the importance of direct licensing revenue continues to grow, he emphasised.

The Guardian has established clear guidelines that it uses to evaluate each licencing opportunity.

Taking on AI’s challenges

Since the emergence of ChatGPT in 2022, The Guardian has been in “a whirlwind of activity” to secure its interests, Hahn said, and that resulted in a three-pronged approach:

1. Legal enforcement. The Guardian makes its content available for personal, non-commercial use only; any other usage requires a license. To protect that standard, the organisation has ramped up legal reviews to identify breaches of terms of service, commercial contracts, and copyright law.

“Clearly there’s a substantial challenge here because … copyright operates in a territorial way. So part of the work has been understanding the jurisdictional differences and the friendliest jurisdictions in terms of enforcement and shifting through those gears of enforcement.”

2. Compliance.  To address vulnerabilities in content access, The Guardian initiated what it calls “Operation Leaky Bucket,” a comprehensive audit of all entry points to its content. This included revising terms of service, reviewing commercial agreements, and, importantly, fighting unauthorized bot and crawler activity.

The organisation now deploys advanced bot detection and user agent auditing, blocking harmful bots that harvest content. This not only protects content but also strengthens the publisher’s negotiation position. “If you can constrict access, you will drive demand,” he said.

3. Commercialisation and deal-making. With protections in place, The Guardian has moved aggressively into licensing deals. It has defined core licensing principles, developed pricing models, and studied AI use cases across the market.

“As we’re doing more and more deals, we are learning from live negotiations,” Hahn said. The organisation is already closing deals, with announced partnerships including OpenAI, Mistral, and ProRata, and has more in the pipeline.

Its partners range from major tech companies like Google and Microsoft to new AI players to legacy aggregators like LexisNexis and Factiva, who are adapting their services to include AI capabilities. These deals vary widely in scope, from simple retrieval-augmented generation (RAG) integrations to broader archive licensing, Hahn explained.

The Guardian has identified a broad range of organisations interested in using its journalism.

“It is not a one-size-fits-all,” he said. “These deals all come in different shapes and sizes in terms of that, the use cases, the value, and so forth.”

Lessons from the licensing frontline

Hahn concluded with some of the key takeaways from The Guardian’s journey so far:

• Understand and organise your datasets: “A key component of deal-making is understanding your available content sets, ensuring they’re stored in structured databases and offering high levels of plug-and-play through APIs.”
• Know your content and rights: Organise archives in structured databases and ensure rights are clear and licensable for AI purposes.
• Protect your assets: Bot blocking and monitoring are essential to prevent misuse and protect negotiating power.
• Understand your customers: Examine how existing clients are using your content and whether they’re complying with current licenses.
• Close the knowledge gap: Learn the technical and commercial realities of AI to negotiate effectively with platforms.
• Establish your licensing principles: Create a framework for what makes a deal worthwhile, including risk appetite.
• Lean into collective action: Collaborate with other publishers and trade associations to gain leverage, share intelligence, and shape regulation.
• Do the deals that are right for you:This will do more than help the individual publisher; it will help “create a functioning market for AI licencing [that] can only help us in terms of our legal rights and court cases.”
• Be flexible and unconventional: In this new and uncertain market, creative deal structures and long-term thinking are key.
• Keep up regulatory reform and pressure: Stay active in coalitions and policy conversations — such as the UK’s creative coalition fighting proposed copyright changes.

By demonstrating a willingness to license and protect its content, The Guardian is helping set the precedent for responsible AI use, and as publishers work together it will provide the “strength in numbers” that will protect and enforce their rights.

“Doing more deals, particularly in search, is going to create a diversity in the search environment which long-term may help publishers in building de-dependency on current monopolies,” Hahn said.

YOU MUST BE A MEMBER TO ACCESS THIS CONTENT

Become An INMA Member or Log In

… … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … . .