Winning Amid Generative AI Disruption: A Proactive Action Plan for Content Owners

The rise of generative artificial intelligence is impacting nearly every industry. Particularly threatened by GenAI are publishers and content owners, which could see their business models upended as GenAI disintermediates their access to audiences, suppresses the attribution of information and eliminates traffic referrals provided by search engines. However, in addition to presenting an existential threat to publishers, GenAI offers many of them a substantial growth opportunity.  

L.E.K. Consulting expects content publishers that “lean in” and take a proactive approach with GenAI to benefit the most, while those that remain bystanders will likely be most negatively impacted, as regulation is failing to keep up with the rapidly evolving technology landscape.  

To secure their position in the future of content creation, distribution and monetization, publishing companies must determine their optimal path forward, develop strategic and tactical plans to realize that vision, and engage with stakeholders across the GenAI ecosystem. GenAI threatens to disintermediate content publishers, taking away attribution and traffic referrals. It also threatens to commoditize content if consumers cannot distinguish the source and/or do not engage with content sources directly. Further, in its current form, GenAI is enabling widespread content “crawling” and distribution without monetization. That said, we see a significant opportunity for content owners to mitigate these risks and gain new methods of monetization in the process. 

The publisher’s conundrum 

Companies that own data and content on which GenAI models are trained, from financial/professional data and solutions providers such as Bloomberg to major consumer publishers such as Hearst and The Washington Post to digital/social platforms such as X (formerly Twitter) and Reddit, face a particularly difficult set of decisions around how to respond to such models. Much of their content has already been ingested and contributed to training GenAI models (see Figure 1). 

Our research and analysis highlight that high-quality content, which many publishers and content owners create, is essential to training GenAI models for three key reasons: 

1. Accuracy: While GenAI training leverages many forms of data, including both formal writing and informal person-to-person exchanges (e.g., Reddit and X data), premium publisher content is paramount to ensuring the accuracy of GenAI outputs. 

2. Quality: Access to high-quality editorial content is required to ensure models are capable of outputting error-free, well-written prose with appropriate grammar, wording and syntax. 

3. Recency: Timely information is necessary for many GenAI use cases, the most prominent of which is AI-enabled search (e.g., Google Bard, Microsoft Bing Chat). 

The upside of GenAI 

Understandably, many content and data owners have concerns that providing access to GenAI platforms could lead to disintermediation of their customer/audience relationships and irretrievably impact revenue sources, including advertising, affiliate marketing and subscriptions. However, content companies should also consider the opportunities that GenAI presents to create value in new areas and develop new revenue streams from existing intellectual property (IP).

While GenAI-enabled search (e.g., Google Bard) and chatbot interfaces (ChatGPT) are the most widely recognized examples of GenAI, we expect the majority of GenAI-driven value creation to take place in business-to-business (B2B) contexts (e.g., Microsoft and OpenAI making GPT4.0 available to Salesforce for use in its business-facing products) and business-to-business-to-consumer (B2B2C) contexts (e.g., Microsoft and Open AI making GPT4.0 available to Mercedes-Benz for use in its consumer-facing products). B2B and B2B2C use cases will vastly outweigh first-party GenAI search and chatbot products in revenue terms and will offer data owners the opportunity to participate in new value creation across a vast array of industry verticals in which they previously had no role (see Figure 2). 

We expect content owners that move quickly and work to package and deliver their data in ways that enhance efficiency and value for GenAI operators will be best positioned to capitalize on the data monetization opportunity. Content and data are both highly valuable to GenAI model training and critical to unlocking broader AI-driven products and services, but there are diminishing returns on the volume of comparable content and data introduced to model training — so being a bystander will have negative consequences. But publishers can take a proactive approach, potentially including collaboration with other publishers. In fact, collaboration may prove essential: Developers indicate they are already able to “weight” high-quality content more heavily (rather than simply adding more) in the training process, creating substitutability. 

As publishers and content owners outside the core GenAI value chain look for ways to capture their “fair share” of the enormous opportunity presented by the rise of GenAI technology, they should consider a range of potential monetization approaches, for example: 

• Negotiated licensing: Given the critical need for premium, reference-level content to train GenAI models, there is a near-term opportunity for publishers to establish frameworks and structures for licensing and compensation. Certain major publishers — including the Associated Press and The New York Times — are already reportedly in talks (or have signed deals) with leading GenAI players to license vast libraries of their content for model training purposes. 

• Consumption-based API access: Other content and data owners, including X and Reddit, have long provided access to their data via publicly available application programming interfaces priced on a self-serve consumption basis rather than taking a negotiated licensing approach. Having recognized an opportunity to monetize their data at a higher rate in the context of GenAI, they have rolled out new pricing models and access controls as they work to capitalize on this opportunity.  

• First-party GenAI model creation: Some data owners, including Bloomberg and Adobe, have thus far elected to keep proprietary data in-house and build their own GenAI models, which are expected to drive enhanced competitiveness and new revenue opportunities over time. Such models currently appear to be most prominent among holders of nonpublic (i.e., inside paywall) data. However, looking ahead, it may be possible to leverage technical, legal, regulatory or contractual requirements to unlock such opportunities for entities (e.g., newspaper publishers, which have long been subject to web scraping) that currently lack full access control over their data, enabling them to license some of their content while keeping other content for the creation of first-party GenAI products or in-house tools. 

Action plan for publishers and content owners 

L.E.K. Consulting’s Media practice has been actively supporting clients in GenAI data monetization and licensing strategy across B2C/B2B publishing and data, as well as across film/TV and music/audio. Our experience suggests content and data owners should pursue five key steps in developing their optimal strategy and approach for monetization in the context of GenAI: 

1. Understand current usage of your IP by GenAI platforms, including the volume and nature of data ingested. 

2. Evaluate expected GenAI market growth by use case to help inform data valuation, in combination with a view of current IP usage by GenAI platforms. 

3. Assess the relative value creation potential of selling/licensing content for model training purposes versus creating your own first-party GenAI products based on the content. In many cases, there may be opportunity to pursue both paths. 

4. To support licensing, assess your optimal pricing, packaging and delivery approach as well as key contract elements and data usage constraints so as to mitigate disintermediation risk. 

5. Establish access control and identify enforcement and risk management strategy/tactics (e.g., watermarking IP, blocking crawlers). 

While the course of GenAI impacts remains uncertain, the likelihood for disruption among publishers and content owners is indisputable. In order to influence outcomes and better realize available opportunities, the time for careful planning and action is now. For a more comprehensive perspective on the opportunity that GenAI presents, publishers are welcome to contact L.E.K. 

L.E.K. Consulting is a registered trademark of L.E.K. Consulting LLC. All other products and brands mentioned in this document are properties of their respective owners. © 2023 L.E.K. Consulting LLC