Report from the Wall Street Journal
In Brief – Meta has been in discussions with several prominent media companies, including Axel Springer, Fox Corp. and News Corp, about licensing their articles for use in its AI-powered products, including a variety of chatbots. Meta announced an AI content-licensing deal with Reuters last October but has only started having discussions with additional large publishers in recent months. Media licensing deals would represent the latest shift in the company’s policies regarding paying for news content, with the company once paying news publishers tens of millions of dollars to include content in its Facebook News tab, but then phasing out publisher payments, and even blocking news posts when payments were mandated by law. The rise of AI chatbots and AI-enabled search tools has resulted in some publishers trying to limit access to their sites by AI crawlers, at least without compensation, and some AI developers, including OpenAI and Amazon reaching licensing deals within major news industry companies.
Context – All AI “training” using media content is not alike. Massive amounts of text content available online was used to train the underlying Large Language Models (LLMs) that run chatbots. Basically, all the text on the internet has been used, including all manner of copyrighted works. A series of copyright lawsuits, many in US courts and brought by authors and major publishers, including media companies, will likely determine if that training violated copyright law. The issue saw a pair of dueling decisions from federal judges in June, one defended LLM training as “fair use” while the other rejected that claim, but the judge defending fair use in training rejected piracy in gathering the text. Anthropic is trying to craft a settlement to distribute $1.5 billion to authors in that case. Meta is facing similar charges. These media payments are different. They appear aimed at keeping Meta chatbots current, such as how AI-based search company Perplexity is proposing to pay media companies. Google is being sued by publisher Penske Media for using content gathered in its search indexing to keep its AI products and search current.