How can journalism survive AI?

“We’re putting all our efforts into maintaining the status quo instead of preparing ourselves for the world we’ll be living in tomorrow”. This remark was made by Meredith Whittaker, technology expert and president of the non-profit messaging app Signal, during her visit to Publix last Thursday. In appearances across the globe, she has been warning people about an enormous concentration of power in the hands of a small group of US tech corporations. But what does it mean – for journalists and media outlets that seek to reach the public with independent content and research – to set aside the status quo and focus on the foreseeable future?

The status quo: a great deal of media organisations and journalists were only just recovering from the blow to their business models inflicted by Google and social media platforms. For example, many have brought in digital subscription or membership models that often rely on a balance between free and paid content. The future: generative AI finds every freely available piece of information on the internet, aggregates it, and distributes it directly, without users having a clear sense of where exactly the information came from.

When Google introduced its new feature AI Overview in May, many users initially welcomed it as step forward. Ask a question and instead of just links, you receive a direct summary of content drawn from various sources. Yet, to editorials that for years now have struggled to assure visibility for their content and their articles, the new function signalled a dramatic exacerbation of an existing problem: content from their own websites is shown elsewhere with users no longer redirected to the original page, and the originators are not even paid for this non-consensual provision of information.

A complaint against Google’s AI Overview has now been lodged with the authorities responsible for compliance with the European Digital Services Act here in Germany. It was filed by a broad alliance of civil society groups, media associations, and companies from the digital economy. Their charge: Google systematically disadvantages journalistic content and thus violates rules on transparency and fairness. That NGOs like AlgorithmWatch and industry associations like VauNet are taking a stand together is the right move. It shows a recognition of the future that is already upon us.

The large language models behind AI applications such as AI Overview, ChatGPT or ClaudeAI rely on content created by journalists, specialist editorials, public interest media and many others. Ultimately, however, it is the tech companies that profit more than anyone else.In an interview with Fachjournalist, I already pointed to this problem and explained why specialist and local media are especially threatened by this development.

What is at stake is the visibility and, thus, the long-term financial sustainability of journalism. We are at a critical juncture: just as publishers had begun to establish successful strategies toward broad dissemination of their content in a digital space dominated by tech giants like Meta and Google, the game has taken yet another turn. This time, it seems that automated AI systems could make journalistic media – as providers of research and contextualisation – basically disappear. ChatGPT, ClaudeAI, and Gemini are becoming, for many, the first port of call in the search for information, yet they take no responsibility for content, and they compensate no-one for the work done to gather and collate it. How does ChatGPT “know” what happened in Lebanon yesterday? That question can only be answered by correspondents from international broadcasters and news agencies. Yet, who will pay for the (expensive) work they do on the ground if the results are free via chatbot?

It is under these circumstances that Meredith Whittaker, president of Signal, took to the stage here at Publix. I asked her a question of concern to many editorials at the moment: what does it mean for democracy if the flow of information is dominated by large AI models? And how should journalism – from the New York Times to Der Spiegel – respond?

Her answer: “We have to start asking ourselves how we can preserve journalism as a force that opposes authoritarian movements, even if the business model completely disappears”.

Whittaker is speaking not only to a risk here, but to a critical juncture. If the dominant business model in journalism – dissemination funded by advertising – crumbles further, we will need a clear new direction: toward direct relationships with the audience instead of click rates, toward spaces for dialogue instead of a narrow focus on distribution, and toward trust garnered by editorials who know the needs of their users, tailoring formats and topics accordingly.

Every day at the Publix building, I speak with media workers who are finding new ways to directly reach their audiences. Correctiv involves citizens in investigative research, the agency Headliner brings journalistic material to theatre stages, and new local media outlets like Loky (Berlin), Tsüri (Zürich) and Bajour (Basel) manage to inspire even 20-year-olds about their city with an accessible newsletter style.

Publix is an innovation centre for common-good-oriented media. Together with the Media Forward Fund, which Publix co-founded, we are actively looking for exactly this type of approach and hope to bring them greater visibility in the coming months. Who out there is building stabile, robust relationships with their audience, landing in their inboxes, their podcast feeds, and on their home pages? We want to gather together best practices, academic knowledge, and models consisting of concrete solutions, thus helping to shape the information eco-system of tomorrow.

In the age of generative AI, journalism needs greater protection, it needs strategic support at this transformative juncture. That is precisely the work we are doing here at Publix.

Fotocredit: Paul Probst