"Generative AI is an Existential Threat to Journalism"

Generative AI scrapes journalistic content without paying for it: This is rapidly intensifying the financial crisis facing journalism. For media organizations, pushing back against powerful tech giants individually while already operating under severe financial pressure presents a daunting challenge.

Media associations can act as intermediaries by uniting their members' interests to negotiate on a larger scale. DW Freedom had the opportunity to speak to the News/Media Alliance, a nonprofit organization that represents over 2,200 news and media publishers in the US.

DW Freedom: With AI-powered chatbots and search engines providing instant, personalized answers, users are less likely to visit websites that host the original source of information. What are the implications of this shift for the media industry?

Danielle Coffey: We see it as an existential threat. Journalistic work involves significant investment in fact checking, vetting, and human curation – elements we believe are essential for producing reliable information. AI products rely heavily on such quality content to deliver accurate answers. This was confirmed by a white paper we released, showing that journalistic content is among the most frequently used by AI systems due to its quality and accuracy. Essentially, these tools depend on the very standards of journalism we uphold, yet they undermine the sustainability of the industry when using our content without authorization and compensation.

Traditional AI models rely on a specific set of data used for training. How does retrieval-augmented generation (RAG) differ from this approach, and what does it mean for publishers?

RAG is particularly concerning for publishers because it pulls real-time, up-to-date content to generate responses. This means it uses today's articles and breaking news that were created through journalists' hard work to uncover and report time-sensitive events. By extracting and repurposing this valuable content for the same audience, it puts us in direct competition, often without offering any recognition or return to the original publishers. There's no business model for us in such a scenario, and I believe this is what makes it an existential crisis for journalism.

Do you see licensing and similar business arrangements between AI companies and publishers as an effective way to deal with the problem?

I think it's a step forward. However, and this is a big "however," there's nothing better than having sustainable legal rights. The current business arrangements don't reflect fair market value because, by definition, a market is suppressed when there is uncertainty. This means that prices are held down because AI companies aren't sure if they are legally obligated to pay.

What's your outlook on the regulatory environment for publishers in the US?

I believe we have a strong case. In the US, case law considers several factors in the fair use test. For example, if AI companies' use of our content is not transformative and serves the same purpose, which it clearly does here, then it's likely not a fair use. Another factor is market harm. That's where the licensing agreements already in place help us because they show that there's a market for our content. If the unauthorized scraping by AI companies is deemed "fair use," that market would vanish. Publishers would never be paid, and a court would have to decide that this market harm – meaning the evisceration of our industry – is acceptable. I think that would be a difficult conclusion for a judge to reach, especially given the constitutional importance of a free press in the US. So, I'm optimistic that the courts will eventually rule in our favor.

What can media associations like the News/Media Alliance do to support publishers in the meantime?

We're already facilitating collective licensing agreements because we can't afford to wait for the courts to decide. By representing a broad range of news and media outlets, we're able to negotiate scalable business arrangements on behalf of our members. Our role is to act as an intermediary, helping publishers understand the diverse interests and needs of AI companies and ensuring fair compensation for their content.

Interview: Elena Köhler