Now is the Time for Premium IP Holders to Develop Licensing Models for Gen AI

Now is the Time for Premium IP Holders to Develop Licensing Models for Gen AI

June 25, 2024

Authors

Phil Schuman, Senior Managing Director
Daniel Punt,
Senior Managing Director
Sumeet Gupta, Senior Managing Director
Russell Simon, Senior Consultant


OpenAI’s recent $250 million deal with News Corp underscores the urgent need for Premium IP holders to forge content licensing partnerships with Generative AI platforms. This agreement grants OpenAI access to extensive archives from major publishers such as The Wall Street Journal and The Daily Telegraph. As more companies, including Getty Images with NVIDIA and Reddit with Google, pursue similar deals, the Gen AI content licensing landscape is rapidly evolving. Premium IP Holders must act now to establish partnerships and set business precedents. With few standardized terms and ongoing legal uncertainties, this is a pivotal moment to maximize the value of their content. FTI Delta helps media companies navigate this new terrain, ensuring they secure a meaningful share of the emerging opportunities in the Gen AI market.

The recent news that OpenAI has struck a deal with News Corp is the latest sign that premium IP holders need to act now to develop content licensing partnerships with reputable, leading Gen AI (generative artificial intelligence) platforms, or risk being left behind.

OpenAI’s multi-year agreement with News Corp is said to be worth more than $250 million over the next five 5 years and gives OpenAI access to current and archived articles from publishers, including The Wall Street Journal and The Daily Telegraph. This comes as an increasing number of premium rights-holders are pursuing Gen AI content licensing deals. There are now notable first-movers for Gen AI content licensing for photos (Getty Images – NVIDIA), online text-based content (Reddit – Google) and many leading publishers (nine9 currently have deals with OpenAI including The Associated Press, LeMonde, News Corp. and Dotdash Meredith). These deals primarily provide content to inform large language models (LLM’s), while publishers receive potentially significant financial rewards and a seat at the table to develop new products and features in partnership with Gen AI platforms.


However, data on Gen AI licensing deals for video and audio remain scarce, and we still are in the “Wild West” when it comes to Gen AI regulation and the business models that will emerge. There is still significant outstanding litigation on IP used for training LLMs, as the New York Times and Alden Global Capital lawsuits against OpenAI remain ongoing. Copyright laws are still evolving and will need to assess both (1) the extent to which AI-assisted assets are copyrightable and (2) to what extend copyright protections should be extended to publicly available IP used to train AI models. We expect innovations in Gen AI to far outpace changes to copyright law, with the U.S. Copyright Office only launching their AI Initiative in 2023; it still has not published results from its first study. Plus, while regulation is frequently associated with restrictions, it often allows mainstream adoption of new technologies, and it is far from certain that future Gen AI regulations would restrict the use of existing content to inform LLM’s.

In FTI Delta’s view, premium IP holders cannot sit on the sidelines and wait for legal issues to resolve before charting a path forward for content licensing to Gen AI platforms. What happened at the outset of the internet serves as a cautionary tale for premium IP holders this time around. At the outset of the internet, technology-driven distributors and those who owned customer-facing products succeeded, but some content owners who sat on their hands while legal frameworks were established to protect IP in the digital realm suffered. Some premium IP holders were unprepared for the rapid shift to digital business models, fearing that they would cannibalize traditional revenue streams from physical media (with publishing and streaming being strong examples). Some premium IP holders lost significant control over their content and revenue streams to new digital players who were quicker to adapt to the changing landscape.

Given that there are still few standard terms and conditions for Gen AI content licensing, this is the time for premium IP holders to set the business precedents with Gen AI platforms; otherwise, they risk being sidelined as Gen AI platforms establish more standardized terms and conditions with partners. In order to get the right set of precedents to allow premium IP holders to take a meaningful share of the value created by their underlying creativity, FTI Delta is advising media companies on how to maximize their IP in content licensing deals. We have developed an initial framework to allow premium IP holders to form a view on what use cases are acceptable in licensing deals, assess the rights Premium IP holders may be prepared to sell, and establish what the price should be in deals in order to achieve an acceptable premium. This is a helpful starting point, as the framework is adjusted based on the specific IP being evaluated.

  • Step 1: Assess early-stage examples as well as analogous licensing models to identify licensing best practices. Existing licensing terms for relevant IP licensed to Gen AI platforms (though there aren’t many) will be helpful, along with existing content licensing structures that are industry specific (i.e., news syndication, photography and image licensing).

  • Step 2: Establish use cases for how the specific Gen AI platforms would likely use IP and establish a view of the financial, unit economics and business KPI’s for how licensed content would impact the specific Gen AI platform. Premium IP holders need to understand how Gen AI companies ascribe value to specific content and how value changes based on the properties of the content that could be licensed. Scarcity is very important from the licensee POV; if there are other sources available for a Gen AI platform to license content with similar unique properties, premium IP holders will have less leverage.

  • Step 3: Once specific IP has been identified, premium IP holders need to effectively value content to drive negotiations, with a full understanding of platform and end-consumer use cases. Because there are not many early-stage examples to use as a baseline, we recommend valuing content based on many factors which will typically include an assessment of the costs to create it and the scarcity of the content in the marketplace. It is also important that premium IP holders are able to get a sense of how much of the provided content would be used to train LLMs.

  • Step 4: Develop the go-to-market plan and negotiate. Premium IP holders and Gen AI platforms need to negotiate over many items including the deal payment structure and key restrictions on the use of IP. There is a wide range of possible deal payment structures as there are potentially many different monetizable events for premium IP holders, including when content is ingested in models and when it is used by end users. There also can be incremental fees for model updates, as models reliant on premium IP frequently need to be refreshed with updated information (i.e., news). Premium IP holders need to protect themselves and can do so by including language that sets up structures for possible royalty allocations to participants who have created content being licensed and by including indemnity and liability language.

By following these steps, premium IP holders can set themselves up for success as they negotiate with Gen AI platforms. Now is the time to maximize value.

The views expressed herein are those of the author(s) and not necessarily the views of FTI Consulting, Inc., its management, its subsidiaries, its affiliates or its other professionals.

Subscribe to Access Our Latest Insights

Get exclusive access to content and market updates of your choice, directly to your inbox.