Mathias Döpfner
President and CEO of Axel Springer
Digitalization manifests itself in three key moments for me: In 1995, I used a newspaper website for the first time – I think it was the “taz” – and it was immediately clear to me: the newspaper as we know it is an outdated model. Why the editorial deadlines, the printing presses, the transport by truck? It was obvious: the idea of the newspaper must be emancipated from paper.
In 2007 I held an iPhone in my hand for the first time and thought: this will be the reading device of the future. As mobile as a newspaper, only much more practical. No black fingers, writing in any size, no waste of paper. And: this is where the future of digital subscriptions comes from, so it’s back to being a business model.
Then, at the end of 2022, ChatGPT. I typed in the first questions, read the answers and knew: most of what we’ve done in journalism in the last few decades will be done by these machines in the future. That will make us redundant or better. What really matters now is back in the spotlight: news, in High Middle German: die Zeitung.
Generative artificial intelligence (AI) is the biggest revolution since the invention of the Internet. Developments such as Bard (Google) or ChatGPT (OpenAI/ Microsoft) resemble a monstrous technological wave that either crashes over publishers and destroys them, or strengthens journalism and raises it to a whole new level. A new level of quality, relevance and commercial appeal. Or the fall. It’s all or nothing.
Great technological innovations often trigger even greater waves of cultural pessimism. When the railroad was invented, people feared the soul would fall by the wayside. When the automobile was invented, the German emperor continued to rely on the horse. When photographs could be printed, some feared for the future of the written word. When computers entered the newsroom, many journalists were sure that the quality of texts would suffer. They could correct more quickly and therefore formulate more carelessly and superficially.
The same is happening now: AI has caused a real panic. In Hollywood, actors and screenwriters are on strike because they fear being replaced by AI. And in some newsrooms, too, there is more skepticism about progress and a desire to be able to slow down innovations. But that never works. And it has always been the surest sign of ruin. The mocking look at banal, ridiculous or just plain wrong results from bots evidences both the loss of reality and insecurity and the defiant self-assurance: humans will always do better. We won’t.
ChatGPT or Bard are answering machines. They have impressively intelligent answers for almost everything. Of course, the information provided must first and foremost be truthful. And it remains the task of journalists to point out where the machines intervene and when they are wrong. But algorithms learn fast. And each user improves their quality. Their few errors are transitory phenomena. Soon, machines around the world will answer based on real-time data, in text, audio and video formats. These new answering machines have the potential to make or break not only search engines, but media as we know it. Why keep reading a newspaper or an app when answering machines instantly give you what you want to know? All available information is aggregated by so-called “big language models” in fractions of a second and processed as needed.
So what matters to journalists in the future is the information that is not yet available. The winners are the ones who discover what should not come to light. They see what no one has seen yet. The competition for exactly that, which has always been the essence of journalism, is open again.
Whether generative AI will empower and strengthen us journalists or make us superfluous depends solely on whether we draw the right conclusions from the technological advances and opportunities.
One thing is clear: we need to understand, embrace and shape the opportunities of artificial intelligence rather than leaving this to the platforms alone. Cultural criticism, progress and technological skepticism as first reflexes are wrong.
But we must be realistic: If at the same time we do not quickly ensure a secure legal framework that gives IP creators a fair share, there will be no business model. And without a business model, there is no competition. And without competition, there is no independence and no quality.
In the following, however, I would like to focus exclusively on opportunities:
No one knows exactly how things will evolve, how habits and products will change. There will probably soon be AI-based editors-in-chief as avatars and video presenters as bots. What is clear is that AI will support many elements of traditional editorial processes and replace them step by step. Text aggregation, i.e. gathering existing information, fact-checking, correcting spelling mistakes, translation, layout, image selection, text editing and production in the classical sense: all this will be done by machines in the future, and sooner than you might think. If we try to block this development in order to protect jobs, we are riding a dead horse and leading ourselves, the media industry, to ruin.
Sports results, election results, stock quotes or labor market data, for example, are already created and analyzed automatically in an excellent way. Soon AI will be implemented deeper into our processes. A few months ago, we already built a print newspaper using AI without deviating from digital article stock. From digital to print in a matter of minutes. This can help produce newspapers and magazines cost-effectively for a little longer than without AI.
But these efficiency gains can help above all to hone the true unique selling points of digital newspaper brands. It’s not about cheapening our offer, it’s about improving it. We need to reallocate financial resources. Or better: invest money in spirit.
Everything that has to do with content aggregation and production will gradually be replaced by AI. Everything to do with original content creation will remain the domain of journalists.
When almost everything changes, it is always especially important to define what will most likely not change, i.e. what to continue to focus on in the future.
I see three things above all else:
A good story is still a good story. The result of an investigation or a reporter’s observations do not lose relevance or interest because they are not read on paper, but on a smartphone screen. A good story doesn’t get worse because you used ChatGPT when researching or checking historical facts. The power of a good story lies in its relevance, its emotional intensity, its entertaining nature and the tension it creates.
Which brings us to the second criterion of timeless quality: A good story must be well told. No matter how important and moving the essence of a story may be, if it is not well told – that is, vivid, concrete, dramatically thought out – it will bore. Language is the decisive matter. Those who love language, handle it with passion and care, treat it with respect, and yet always playfully challenge it, have an eternal competitive advantage.
To exaggerate, one could resort to the anecdote of the old married couple at the café: the two are sitting at the table, drinking a brandy, and every few minutes one of them says a number out loud, whereupon they both burst out laughing. At one point, a young man joins them and asks them politely what they are doing. It’s very simple, they explain. We’ve known each other for decades, we go to the café every Sunday and we always tell each other the same jokes. At some point we started giving a number to each joke instead of tediously repeating them. I see,” says the young man, and continues to watch the two of them. 26. laughter. 35. loud laughter. At some point the observer perks up and tries it in turn: 17. no one laughs. Iron silence. wasn’t the joke funny? It was, says the woman, but you have to know how to tell it well.
Which brings us to the third rule. Laugh rather than bore. A good story should not only move its readers, sadden them or make them reflect, it should entertain them as often as possible. There is no better and deeper link between a text or a movie and the laughter of the reader or viewer.
AI can’t do all that on its own, but it can help with almost anything. And when it comes to defining the limits of AI in the process, we should be wary of rash certainties. I, for one, have had to correct such challenging paradigms all too often in recent years.
Some examples: AI doesn’t have feelings? Not true. If anything, AI has long been able to simulate feelings quite well and even situationally. And as for authenticity, please don’t be arrogant: feelings are sometimes simulated as mechanically by humans as by machines.
AI has no creativity? No, it doesn’t. Anyone who has dealt with machine-made music and algorithmically created poems or images should confirm otherwise. Some of the results are still clumsy, others are already phenomenally inspiring.
And finally: AI has no sense of humor? That’s not true either. Even in the remotest realm of nuance and irony, well-trained “great language models” can occasionally rise to the occasion. Whether text or voice, bot conversation not only offers access to an endless pool of jokes, but also many spontaneous punch lines.
So instead of pitting the good human against the bad machine, it would be wiser to use AI to improve human journalism. The bot as a service provider. As an assistant to journalists, making us faster and better at creative thinking, at developing original and relevant topics and forms of presentation, at maverick opinion pieces and intelligent analysis, reports based on real experiences. Whether it’s expertise conveyed in a comprehensible way, entertainment or live experiences, AI is everywhere and allows us to focus on the essentials.
Most important is still investigative journalism: the in-depth research over days, weeks and months that brings new facts and perspectives to light. While AI streamlines the material and quantitative aspects of the profession, it can help focus on the intangible and substantive intellectual essence of journalism.
This is good news for any passionate journalist. The creation of exclusive content is increasingly crucial for the success of publishers. And this means concretely: the most important people in a media hierarchy are therefore the authors.
This development requires a fundamental redistribution of resources and money. A shift from production to creation. Intellectual, creative and journalistic talent is, once again, the most important success factor of a publishing house. This has been forgotten here and there, but it has always been the case.
What is new is a greater focus on technology. Technology is not a secondary element of our business. It is part of the core. Only with the best technology and product developers and the most experienced artificial intelligence experts will publishers be able to achieve their goals. Technologists must be on par with publishers. They must be an early and integral part of any form of product development.
And something else will change: IQ will become less important; emotional quotient will become more so.
While the amount of knowledge has grown exponentially in recent decades, digital technology has helped us access and process ever-increasing and unwieldy amounts. We have become accustomed to outsourcing much of our knowledge. We no longer need to have everything at our fingertips. The expanded recollections of our memory are called Google or Wikipedia. Why memorize everything when it can be searched and found quickly.
This trend is reinforced, accelerated and amplified by generative AI. Even the conclusions of accumulated knowledge can be successfully delegated to some extent – why analyze it yourself when ChatGPT can do it faster and better? But if machines take over more and more frequently and successfully the functions that define part of our IQ – for example, analytical skills – the sheer wisdom of a person’s IQ will be less and less of a competitive advantage. Because anyone can make use of it. IQ can be bought in the future.
This applies less to what we call Emotional Intelligence: emotional intelligence, social intelligence, intuition, creative disruption. Emotional Intelligence is becoming more and more important. The more precise perfect machines become, the more relevant the charisma of the imperfect human being becomes. Emotional leadership and motivation, intuitive creativity and networking make the difference. This is where human intelligence will excel. This is where the competitive advantages and value creation of the future will emerge. This applies as much to research and science as it does to art and culture, to entrepreneurship in general and journalism in particular.
The epochal change that generative AI represents is a historic opportunity. For society and for journalism. If we get it right, journalism will rise from the ashes like a phoenix. If we defend the old structures, soon there will be nothing left but ashes.
If we get it right, machines will be at the service of people. Not people at the service of machines.
Ideally, publishers should return to their spiritual and content essence. The transformation of the music industry is repeated in journalism: the product, there first a record, then a CD, finally a chip, here first paper, then a screen, then a chip and its own retina, i.e. the medium, becomes smaller and smaller, less and less important, finally invisible. Thanks to digitization and AI, the newspaper and its production become dematerialized.
In short: pure content. The news. In old German: die Zeitung. Artificial intelligence then promotes something like the rebirth of the original idea of journalism. Or the rebirth of newspapers.
© Published in Die Welt newspaper / All rights reserved
