Axel Springer CEO on OpenAI deal: 'We must protect intellectual property of news content'

Credit: Robert Downs/INMA

Last year, Axel Springer struck the world's first global partnership with Microsoft's OpenAI platform, which will see the AI company pay the German news group for content to train its large-language models (LLMs).

As part of the agreement, ChatGPT users also receive summaries of news content from Axel Springer's media brands including Politico, Business Insider, Bild and Welt.

This contrasts with The New York Times (NYT) which is suing OpenAI for using "millions of their copyrighted articles" to train chatbots.

Mathias Döpfner, CEO of Axel Springer said last week (24 April 2024) at the INMA World Congress in London that he welcomes both approaches but his gut feeling is that the news industry is on the precipice of its third defining moment within his career. The first was in 1995 when newspapers could go digital, and the second was in 2007 with the arrival of the iPhone. In 2023, generative AI brings a new generation of answering machines "with the potential to completely destroy our businesses as we know them or bring them to the next level."

To be clear, Axel Springer has long been working to reform European copyright laws so that news publishers can claim a revenue share of content used by other companies for commercial purposes. The arrival of generative AI had initially undone about two decades of work, according to Döpfner. But he managed to hammer out a deal with OpenAI in just two months of "very encouraging" negotiations.

He said that - like the internet and the smartphone - generative AI technology is here to stay and can improve many journalistic processes from detecting errors to aggregating content. It is in the collective interest of the news industry, he added, to produce a legal framework that protects the intellectual property of news publishers. That, as a principle, outweighs short-term money on the table.

That legal framework could be new copyright law or a different legal instrument altogether, those details are not yet clear. But it is one of three conditions in his mind for news publishers to not be destroyed by AI. The other two are not to be driven by AI's limitations, and to get newsrooms out of their "political camps". Audiences will need truly independent, non-partisan news more than ever in a world of generative AI.

Joseph Kahn, executive editor of NYT who also spoke at the event, echoed similar sentiments about the need for independent journalism. He said the lawsuit of OpenAI is not hindering "robust coverage" of how generative AI is developing, nor the adoption of such tools internally. Tech partnerships are kept entirely separate from the newsroom.

He is "cautiously optimistic" about using tools for tasks like navigating the CMS or data visualisation. Synthetic voice and text-to-speech are also "excellent" and are being looked at seriously. Translation tools - especially to Spanish - are a top priority, but the tools are not accurate enough yet.

Humans, though, will make all the difference in times of automation. We are a long way off relying on AI to hold "a master of media manipulation" like Donald Trump accountable. Kahn singled out reporters like Maggie Haberman, Jonathan Swan and Michael Bender. They have the journalistic skills and instincts to provide real-time fact-checking, cut through the "conspiracy theories" and "soliloquies" and find the right balance of exposure.

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).