How do we deal with generative artificial intelligence that produces reports that are barely distinguishable for readers from those of human journalists? An AI that can curate faster than a well-staffed newsroom and spits out a first draft of an article within seconds instead of hours? Media organisations around the world have been grappling with these questions for months without having any answers yet. However, it seems inevitable that generative artificial intelligence will bring the next revolution to the media landscape.

"This is the most rapid and most fundamental change in the 30 years I've been working in the digital industry," explained Ross Sleight, Chief Strategy Officer EMEA at software developer CI&T at the Mx3 AI media conference in London in December. He assumes that the disruption of business models through generative AI will happen in several, rapidly successive stages. The first phase, which the industry is already going through, is characterised by an enormous acceleration of content production, processes, etc. with the help of generative AI. In the second phase, which is also just around the corner, there will be a hyper-personalisation of content presentation. AI makes it possible to take the wave of personalisation started by social media to the extreme.

What sources will the new gatekeepers draw on and how can media professionals earn money if AI provides answers based on their content? Axel Springer was one of the first publishers to take the lead here and entered into a partnership with OpenAI. The Associated Press had already agreed a cooperation with OpenAI in the summer. However, the agreement with Springer goes even further. The deal that was concluded with the Associated Press only covers large parts of the archive that was made accessible for OpenAI.

Springer, meanwhile, is the first company in the media industry to be paid by OpenAI for the use of its own current content to answer questions from ChatGPT users as part of a subscription model. "The partnership we have entered into with OpenAI represents a paradigm shift," says Springer Group spokesperson Adib Sisani. "The agreement is groundbreaking because it puts the relationship between an AI company and a media company on a completely new footing for the first time. We have granted OpenAI the right to use our existing content for training purposes to feed ChatGPT in return for payment of a certain amount." This relates to publications such as "Welt", "Politico", "Bild" and "Business Insider".

"What's also completely new is the right for OpenAI to use our new content for ChatGPT to answer users' current questions," says Sisani. "This is unique between AI and media companies." For example, if ChatGPT were asked about the state of the traffic light coalition, ChatGPT could summarise the latest article by Welt author Robin Alexander on this topic and provide a link as an answer.

OpenAI pays a certain amount for this each year. "When Google and Facebook were new, the media industry failed to tap into the platforms as a new source of revenue. We shouldn't make that mistake again this time. That's why we are now negotiating with almost all AI companies in Europe and the USA," says Springer Group spokesperson Adib Sisani.

According to reports, Springer is pocketing tens of millions for this non-exclusive deal. For a small publishing house, that may sound like a lot of money. For Axel Springer, however, it is at best a nice bonus. In the first half of the year alone, the media group generated revenues of just under 1.9 billion euros. Should there be a disruption in media usage behaviour, significantly more revenue is at stake than can be recouped here.

Springer's largest shareholder is the US financial investor KKR with 35.6%. The private equity giant also senses lucrative AI business in other fields: "In general, we see great potential for generative AI to create value in a variety of industries - for example, in supporting business processes and improving digital and technological products and the way customers interact with them," says Franziska Kayser, Partner at KKR.

Many media managers are currently still cautious about the Springer deal. In the "Financial Times", one media executive explained that monetisation works best on one's own website. If the consumption of quality journalism were to take place on another platform in the long term, this would certainly not be a good thing.

One problem with the Springer deal, which some media analysts categorise as potentially groundbreaking, is the gatekeeper function it grants Open AI. Anyone who asks ChatGPT today for an assessment of a particular medium on a topic will receive the answer that the AI "unfortunately does not have access to the medium's website". So in future, the platform will only be available to those who make a deal with the AI start-up. This immediately raises questions about diversity of opinion and freedom of the press, should AI services such as ChatGPT actually replace search engines and news sites as the primary source of information for many people in the future.

In view of the expectation of enormous increases in revenue in the coming years, every media company feels obliged to invest in artificial intelligence. However, not every project is thought through to the end. According to media observers, one problem with many AI projects is that they are approached as pure technology studies without a business strategy. "There has to be a business need and an operational goal for the use of AI. It doesn't make sense to reorganise your business around AI alone," said Matt D'Cruz, Partner at HR consultancy Martin Tripp Associates.

The still recent history of the use of generative AI in the media has already shown how quickly ethical and journalistic standards can be overturned. Around a fortnight ago, Ross Levinsohn, CEO of publishing house The Arena Group, was sacked without notice after it emerged that the traditional publication "Sports Illustrated" had not only published AI-generated articles. They were also attributed to AI-generated authors as if they were real people - including a fictitious personal history. Anything that works is far from legitimate.

Other media companies are also pushing the boundaries of technology. G/O Media, which owns the tech website "Gizmodo", has already published several purely AI-generated articles without journalistic involvement. Although numerous errors have been identified, editorial director Merrill Brown has announced that they will continue to do so in the future.