Big tech companies OpenAI, Google, and Microsoft are in discussion with leading publishers concerning journalism and generative AI. The biggest issue in their negotiations is copyright, as the companies are looking to use news content to train AI models and generate content for chatbots, and news providers want to be compensated. Some publishers, including News Corp, Axel Springer, The New York Times, and The Guardian, have talked with at least one of the top AI firms, discussing a subscription model with one of the companies that would allow publishers to use the content for technology. The tech companies are willing to pay millions, with some interested in discussing how they can help publishers generate revenue.
The negotiations remain in the early phases, with copyright-compliant use of news content for AI training estimated between $5 million and $20 million per year. The Financial Times reports that Axel Springer CEO Mathias Döpfner proposes a quantitative model comparable to music streaming and could put smaller regional and local news outlets at a disadvantage. He proposes an industry-wide, collaborative solution.
Google has reportedly intimated an interest in licensing news content for AI training, has devised a model for News UK and The Guardian, and is in talks with European and US news outlets. Google has confirmed the talks but has officially declined to comment on the status or content of discussions. Another possibility entails giving publishers more control in allowing or disallowing content use.
OpenAI CEO Sam Altman reportedly has spoken with News Corp and The New York Times. OpenAI has launched ChatGPT with internet access, acknowledging that this is an innovative way to interact with the internet, promising to add value to users, and is open to ideas on how to return traffic and contribute to the ecosystem’s health.
The issue with copyright in artificial intelligence and publishing surrounds two implications. The first relates to the journalistic content that forms part of the training data, and the second issue concerns internet-connected chatbots that access journalistic content in real-time to generate responses. The creator of the unique content, the publisher, is the one who often gets nothing. Web traffic declines, resulting in lost branding.
Artificial Intelligence and publishing affect all content providers from video content and podcasts, or Spotify summaries, audio clips, or shortened videos, that may end up as content sources for the AI-generated chatbots. The most affected content types are news, reviews, and how-to content.
