Altman Solon is the largest global strategy consulting firm exclusively working in the TMT sectors. We examine how digital media publishers should approach the rise of generative AI. Should they pursue exclusive licensing deals with generative AI platforms? Should they incorporate generative AI products into their properties?
Generative AI: The biggest tech story of 2023
The rise of generative AI is the biggest tech story of 2023. Generative AI is a subset of artificial intelligence that can create new and original content based on a user’s prompt such as text, images, video, and audio.
By the end of this year, S&P projects that generative AI products will earn $3.7 billion in revenue. In fact, we found that in the business world, generative AI tools have the potential to improve productivity across numerous business functions, including (but not limited to) software development, marketing, customer service, and product management.
In contrast to generative AI, 2023 has not been kind to the media:
- 2023 has seen a record number of media layoffs, with over 17,000 jobs cut in the U.S. alone
- Advertising CPMs1 are in decline in North America
- Promising startups, like the once-$5.7 billion-valued Vice Media and the critically acclaimed BuzzFeed News, have declared bankruptcy or been shut down.
How generative AI will change digital publishing is yet to be determined, but publishers are attuned to how it might disrupt their business. Concerns range from worries that emerging generative AI-powered search engines will drastically reduce traffic to their sites, further threatening ad-based models, to concerns over job displacement and intellectual property misuse.
Generative AI chatbots’ global popularity, like that of ChatGPT, stems partly from training on internet content, possibly including paywalled media sites. This contributes to the chatbots’ conversational and seemingly knowledgeable abilities. In response, publishers are claiming copyright infringement.
It’s not all doom and gloom, though. Publishers are seeking compensation from generative AI platforms that use their proprietary data. They aim to steer clear of errors made at the dawn of the internet when content was freely available. They are in talks with the AI divisions of big tech firms to hammer out deals for paid content-sharing agreements.
Implications for Digital Publisher-Generative AI Relationships
Generative AI could be a positive for publishers. We believe that digital publishers should secure exclusive content licensing deals with generative AI platforms. They can also make money by selling their content exclusively to these platforms and experimenting with their own AI-based products.
AI: Powering media organizations & publishers from behind-the-scenes
“We want to send some of our top executives….into the knowledge centers of artificial intelligence in order to become evangelists and change agents.”
— Mathias Döpfner, CEO of Axel Springer —
Despite fears of replacing journalists and other content creators, the industry has used AI and machine learning for over a decade now. These tools promise to increase productivity, automate tasks, and cut costs. They cover a wide range of use cases across different production areas, including:
- Automated SEO keyword tagging,
- Personalized content recommendation, and
- Automated video and image production.
Controversially, media groups have already started using AI tools to cut costs associated with content creation. In some cases, they are partially or wholly outsourcing content creation to AI applications. They are experiencing in real-time the limitations of this approach.
The financial publication CNET issued lengthy corrections for a series of SEO-optimized “explainer” articles written by generative AI tools. They were riddled with factual errors and largely plagiarized. However, AI tools are often used internally. Newsrooms can now integrate AI into their content channels by developing AI products for end readers, calling on publishers for quality control.
Licensing content to generative AI platforms
“[Media]’s collective IP is under threat and for which we should argue vociferously for compensation.”
— Robert Thomson, Chief Executive at News Corp —
Licensing content to large language model (LLM) developers is a direct way for media and content providers to monetize and leverage proprietary data. In July, the Associated Press and OpenAI entered a licensing agreement. It gives OpenAI access to the AP’s archives, dating back to 1985, to help train their model.
In return, the AP will leverage OpenAI to integrate generative AI into its products. While the companies have not revealed the deal’s financial details, an agreement like this is advantageous for OpenAI. With access to the AP’s breaking news, they can improve their developing search capabilities with protection from potential copyright violations.
For the AP, the deal allows them to put a price tag on their content. It also opens the door to exploring additional generative AI-based revenue streams in partnership with OpenAI. In what appears to be a sea of change, sites like Reddit and Twitter have updated the conditions of their APIs. They now require third-party developers to pay for data used to prevent AI companies from scraping their data for free.
Packaging AI products into content
“In 2023, you’ll see AI-inspired content move from an R&D stage to part of our core business, enhancing the quiz experience, informing our brainstorming, and personalizing our content for our audience.”
— Jonah Peretti, CEO of BuzzFeed —
Although still in its infancy, publishers are launching user-facing AI products. These products, trained on the publisher’s proprietary data, are being marketed to users through a subscription or ad-based model.
In May 2023, Skift, the travel industry news and market research site, debuted “Ask Skift,” a proprietary generative AI chatbot. It answers business questions about the travel industry. Ask Skift is built on GPT 3.5 and trained on 11 years of Skift’s archives as well as publicly available data (SEC filings and annual and quarterly reports) from all U.S. public travel companies.
Users can ask three questions per month, and subscribers paying for the SkiftPro product have unlimited access. Skift CEO Rafat Ali says these generative AI tools can foster a more engaged, intimate relationship with end users. The company plans to build out additional features in the future.
“Until now, we had a package-based relationship with our reader….in the form of a story, a research report, or a conference. Now, with new generative AI tech, it’s possible to have a Q&A-based relationship with our users, at scale, in a way that just wasn’t feasible before.”
— Rafat Ali, CEO of Skift —
These generative AI integrations aren’t limited to industry publications. Lifestyle and entertainment publishers are experimenting with them as well.
In February, BuzzFeed launched Infinity Quizzes, which leverages a chatbot built on OpenAI’s API to provide personalized quiz results based on users’ text input. The platform signed gardening products manufacturer Scott’s MiracleGro as an exclusive advertiser, creating a new stream of ad revenues for BuzzFeed. Since launching, BuzzFeed reports that readers spend 40% more time taking AI-generated quizzes than human-generated ones. The company has plans to integrate AI-powered technology into its Tasty app.
While it’s still too early to tell if these generative AI products will be long-term successes, they have the potential to grow user engagement and add value for readers and advertisers. Publishers also benefit from user data that can further inform content and product strategy. However, content providers need to be transparent about data usage and privacy. They’ll also need to invest in AI, tech resources, and personnel to develop these apps.
Keys to a holistic AI content monetization strategy
There is no one-size-fits-all approach for publishers as they try to make generative AI work for their bottom lines. However, we have worked with organizations to ensure they take the following areas into consideration when implementing an AI content monetization strategy…
Complete the form below to discover 3 key considerations for building an effective AI-content monetization strategy.
1 CPM (Cost per Thousand or Cost Per Mille) is a marketing term to denote the price of 1,000 advertisement impressions on one webpage.