In a move that navigates the delicate balance between AI model training and copyright concerns, OpenAI has inked a deal with Axel Springer, the Berlin-based media company behind publications like Business Insider and Politico. OpenAI's agreement with Axel Springer allows the training of its generative AI models on the publisher's content, integrating recent articles into OpenAI's viral AI-powered chatbot, ChatGPT. This collaboration follows OpenAI's earlier licensing arrangement with The Associated Press for model training.

ChatGPT users can now expect summaries of "selected" articles from Axel Springer's publications, even those typically behind paywalls. The snippets will include proper attribution and links to the full articles. In exchange for this access, Axel Springer will receive payments from OpenAI, although the specifics of the financial arrangement remain undisclosed. The deal, spanning several years, does not commit either party to exclusivity, but Axel Springer expresses its commitment to supporting AI-driven ventures that leverage OpenAI's technology.

Axel Springer CEO Mathias Döpfner sees this partnership as groundbreaking, emphasizing the exploration of AI-empowered journalism to elevate quality, societal relevance, and the business model of journalism to new heights. However, this collaboration between generative AI vendors and publishers is not without controversy. Publishers, wary of copyright infringement, have raised concerns about generative models affecting traffic and potentially cannibalizing their content. Google's SGE, a generative AI-powered search experience, has faced criticism for pushing traditional search results down the page, impacting traffic significantly.

The broader context involves publishers grappling with AI vendors over compensation for training models on their content. The media industry, represented by organizations such as Getty Images, The Associated Press, and others, has called for more transparency and copyright protection in AI practices. They argue that current practices not only violate copyright law but also undermine the core business models of the media industry, leading to a reduction in diversity and financial viability. The industry's open letter in August urged policymakers to consider regulations promoting transparency in training datasets and allowing media companies to negotiate with AI model operators.