AI and Content Licensing: A Paradigm Shift or Earthquake? 

We’ll See.

by Frank Bilotto

Dotdash Meredith’s recent partnership with OpenAI has sent ripples through the publishing industry. To quote Marie Griffin, this deal is so “big, big, big,” that we at CLI have already committed to dedicate the next three issues to it. 

To get things started, the Dotdash Meredith-Open AI agreement – coupled with similar deals between AI and media giants like Associated Press, Axel Springer, and Financial Times – makes my prediction of more than a year ago that an AI company will become the world’s largest publisher in the not too distant future seems not only possible, but likely.   

How did we get here?

Dotdash Meredith is the world’s second-largest publisher and operates more than 40 premium publishing sites, including People, Better Homes & Gardens, Verywell, InStyle and Investopedia. Its content includes real-time news, verified health and financial information, and product reviews.

Dotdash Meredith’s parent, IAC, led an effort to unify the publishing industry for copyright protection against the challenges posed by AI’s large language models. I had high expectations, but IAC’s effort ultimately collapsed amid conflicting business incentives among publishers.

Daniel Schaible, of CLI, knows that the struggle to convince publishers to collaborate is as difficult as “herding cats”.  Consistent with John Nash’s Nobel prize winning contributions to Game Theory, I am certain that major publishers acting alone will be detrimental to all publishers in the long run. Schaible predicts: “A fragmented industry plays into the hands of platforms, not publishers.” 

We’ll see.

Essence of the deal

The Dotdash Meredith-Open AI agreement calls for attribution and links from the AI generated content to Dotdash Meredith. In addition, Dotdash Meredith will benefit from the agreement by getting paid for the use of its published material. The deal will give ChatGPT’s 100 million users access to content from the more than 40 media brands. As part of the multi-year deal, OpenAI will display content and links attributed to Dotdash Meredith’s websites in relevant responses to ChatGPT user queries.

Be careful what you wish for

With AI, contractual controls on distribution are irrelevant, and there is no doubt in my mind that Dotdash Meredith missed the mark on requiring accreditation and links as part of the Open AI deal. I think those deal points are meaningless. I remember when I used to tell my kids that they couldn’t trust Google search results, but that they had to consider the source of the authority of the information. A few years later, in 2008, I launched an online medical library designed for physicians, which included more than 2,000 medical journals that sit behind paywalls. What we learned in three years was that when doing medical research, despite having access to authoritative content, 83% of physicians went to Google first. Even for professionals in all fields, who have access to premium content, Google has become the trusted source. Remember that I told you this first, in time, like Google, AI generated content will become the trusted source, as users will not feel the need to check the validity of the answers they receive. 

We’ll see.

So, where’s the beef?

For me, when a publisher is negotiating a deal with an AI company, it comes down to two variables: the shelf life of the content, and the money.  Simply put: the longer the shelf life, the more money it’s going to take for a publisher to license content to AI technologies. Forget about controlling distribution and receiving accreditation for links to your content. It complicates the negotiation and reduces the monetary value of your license and provides you with no added benefit. For many, generative AI already is the trusted source. 

CLI works with publishers who believe their content has a shelf life of anywhere between six months and five years. For content with shorter shelf lives, the publisher should get all they can up front, and be grateful for the additional revenue. For content with a long shelf life, see my article referenced above at the CLI website to learn about how we structure deals that fairly compensate publishers. 

In the near future, if I structure the deals, I am certain that publishers will generate more revenue from licensing their content to AI companies than from their subscriptions or ad revenue.

We’ll see.

If I were king (of all media)

Thirty years ago, individuals had to rely on academic and corporate librarians to do their online research through online information companies like Lexis and Dialog. Sometime in the mid-1990’s students and professionals suddenly had a magic box on their desk that enabled them to do their own research with the entry of a word or two into a search box.  In 1998, I tried to convince senior executives at Dialog to put up a web crawler and mix in what was significantly inferior and publicly available content with the authoritative content of our 300,000 publishers. My novel idea was quickly dismissed as naive. I still believe that if Dialog had put up a web crawler, we would never have heard of Google, which at the time indexed no content of any value. Instead, Dialog and the other major aggregators attempted to discredit the internet search engines and, in doing so, forever lost control of online information. Just as their longtime mission statement declared, Google has organized the world’s information.

Today, publishers are at the same point with AI as librarians were with the internet. Publishers are best positioned to be the dominant creators and distributors of AI-generated content. The publishers own the content that Open AI and other AI companies need for machine learning. Instead of licensing content to AI companies, Publishers should be licensing AI technologies for their own use. But I am seeing it happen all over again. The players who are best positioned to win don’t have the vision or leadership to do it. Unfortunately, by doing deals with AI companies, publishers are missing the opportunity that could keep them relevant in the 21st Century. I am hopeful, however, that before this decade is out (I always liked JFK’s ability to turn a phrase) a visionary major publisher will be utilizing AI to generate articles, and even books. 

In perhaps the greatest miscalculation since Kodak declared digital photography to be “just a fad”, Dotdash Meredith CEO, Neil Vogel, when speaking of his own company’s views on AI, declared: “We will never have an article written by a machine.” Never? 

We’ll see.

Next issue: How the DotDash Meredith Deal is an Existential Threat to Small Publishers.