With AI, news publishers are getting the short end of the stick

More and more large news publishers are striking multi-year deals with AI companies over the use of their content. The future of news lies in the hands of a gerontocratic management class that is securing their retirement rather than protecting the future of journalism.

At the end of August, Condé Nast signed a multi-year partnership concerning the display of its content in OpenAI products. The deal allows OpenAI to display content from Condé Nast brands in its products, including ChatGPT and its AI-driven search engine prototype, SearchGPT. Condé Nast is just the latest in a long list of publishing houses that have signed deals with AI companies, including:

The Wall Street Journal, Barron’s, MarketWatch, Investor’s Business Daily, FN, and the New York Post - the Atlantic, Fortune Wordpress, Vox media, The Texas Tribune, Reuters, Reddit in the US; The Times, The Sunday Times and The Sun in the UK; and the Herald Sun, The Australian, news.com.au, The Daily Telegraph, The Courier Mail, The Advertiser in Australia; The Financial Times, Axel Springer, Time, Dotdash Meredith - with dozens of publications including People, Instyle and Investopedia - Fortune, Le Monde in France, Prisa Media - El País, Cinco Días, As and El Huffpost - in Spain, Il Corriere della Sera in Italy, and many more. 

As you might expect, the terms of these agreements are kept confidential, and reading through the various press releases won’t give you a real sense of what these deals actually entail.The fairy tale that publishing executives are telling us is that these deals will allow people to find trustworthy information on GenAI models. In return, GenAI companies will assist publishers in developing new products and features.

R-r-right. 

Let’s call a spade a spade. The value exchange in these deals is very simple: GenAI companies are securing vital content licensing deals to train their insatiable large language models (LLMs) and avoid costly lawsuits in the process. In return, large publishing groups are desperate to cash in on whatever sum of money they can to offset their faltering business models. 

At first glance, it sounds like a win-win scenario, except for one very important detail: GenAI companies simply cannot survive without being continuously fed an immense volume of high-quality, fresh data. OpenAI has built a $100 billion company by training its LLMs with scraped content from the web (and other sources) without paying a dime for it. They did this knowing that the time and cost involved in securing content licensing agreements would be greater than the risk of copyright infringement lawsuits. And that’s precisely what happened, with the New York Times and a few other media outlets suing them over the past few months.

Now, the real issue is that, according to EpochTech, a research institute, GenAI companies are using data faster than it is being produced, and they could run out of it as early as 2026. This means they don’t have a choice: they must license as much high-quality content as they can to improve their models and, possibly, they need to do it in compliance with copyright law. In this new context, published books and articles written and edited by professionals are the most valuable assets available.

If news publishers were to act as a unified body in negotiations with GenAI companies, the value they could extract from content licensing deals would be immense. Looking at the numbers from a few recent deals, it’s clear that’s not the case. Once again, news publishers are focused solely on improving their individual bottom lines with whatever money they can get, undermining their ability to extract more value.

Take News Corp as an example. Their agreement with OpenAI is reported to be worth more than $250 million over the next five years and gives OpenAI access to current and archived articles from some of the most prestigious publications in the world, including The Wall Street Journal and The Daily Telegraph. While $250 million may sound like a lot, it’s actually quite low when you consider that, with a single deal, OpenAI is gaining access to millions of articles written by some of the best journalists in history, content that will continue to inform their LLMs going forward. By signing the deal, News Corp, along with dozens of other publishers doing the same, has given up its most valuable asset at a bargain price.

So, why would news publishers do that? The answer is simpler than you might think.

Imagine yourself in the shoes of a 60+ year-old executive leading a top news corporation, just a few years short of retirement. Over the last 20 years, you’ve witnessed the progressive destruction of the print business, and the promise that online revenues would compensate for those losses has never materialized. With this in mind, agreeing to a subpar deal to fix your P&L for the next three to five years and secure your own retirement is by far the easiest decision. And this is precisely what most executives are doing.

Like a farmer selling off fertile land for a quick profit, unaware that the soil is the lifeblood of future harvests, news corporations are trading the intellectual capital of their best journalists for a negligible return. When these content licensing deals come to their natural end, GenAI companies will be fully capable of writing stories without their help, using synthetic data to continue their training. At that point, news corporations that signed these deals will face the same, if not worse, financial struggles they had before. Ultimately, years of poor management cannot be resolved with a content licensing deal.

Looking at the glass half full, GenAI companies aren’t interested in producing news content. They understand that journalism is not just about information; it’s about the connection between writer and reader, between human experience and understanding. Content licensing deals are just a way to secure curated data so models can respond to queries like a human would. News publishers simply happen to be the best source of that type of content.

The silver lining is that, despite themselves, the future of news won’t be defined by a group of media executives who can’t see beyond the end of their nose. When the dust settles and the content well runs dry, people will return to what resonates with them on a fundamental level: the authenticity that only human content can offer.

It’s unfortunate that GenAI companies get to dictate the rules of the content licensing game, considering they cannot survive without professional content. It’s equally unfortunate that, once again, the news industry is getting the short end of the stick due to its fragmentation and, frankly, incompetence. Pricing content licensing deals at their true worth is the last chance of survival for news corporations that desperately need resources to evolve into modern media companies. Yet, it seems they are even missing this final opportunity.