Think of any topic loosely related to parenting you can imagine, and you’re likely to find a post about it on Mumsnet, the long-running, wildly popular, controversial British forum for mothers. Over its two-decade history, Mumsnet has amassed an archive of more than six billion words written by its highly engaged users on topics ranging from filthy nappies to indolent husbands. (Not to mention crazy talk about dolphins.)
This spring, after Mumsnet discovered AI companies were scraping its data, the company says it decided to try to strike licensing deals with some of the industry’s biggest players, including OpenAI, which initially said it was willing to explore a deal after Mumsnet first approached it. After talks with OpenAI fell through, Mumsnet announced its intention in July take legal action.
According to Mumsnet, during those early talks, OpenAI’s strategic partnership leader told the company that datasets of more than 1 billion words were of interest to the AI giant. Mumsnet executives were excited. “We spent a lot of time back and forth with them,” Mumsnet founder and CEO Justine Roberts tells WIRED. “We had to sign a bunch of non-disclosure agreements, and they wanted a lot of information from us.”
But more than a month later, OpenAI informed Mumsnet that the company was no longer interested in the partnership, according to an email exchange reviewed by WIRED. When asked why, an OpenAI employee described Mumsnet’s $6 billion dataset as too diminutive to justify a licensing deal, Roberts says. They also noted that OpenAI is primarily interested in huge datasets that the public no longer has access to online, and that it wants datasets that encompass the breadth of human experience.
The company echoed that sentiment when asked for comment by WIRED. “We’re committed to partnerships on big data that reflect human society, and we’re not committed to partnerships solely on publicly available information,” says OpenAI spokeswoman Kayla Wood. “We support publisher and creator choice by offering ways for them to express their preferences for how their sites and content interact with AI in search results and by training generative AI baseline models.”
Roberts says she was “annoyed” by this development. She recalls that OpenAI initially seemed particularly interested in Mumsnet because of the content on the platform, which was largely written by women. “It’s very high-quality conversational data,” she says. “It’s 90 percent women talking, which is quite unusual.”
Over the past year, OpenAI has entered into a number of data licensing agreements with media outlets and platforms, entering into agreements with Vox Media, this AtlanticAxel Springer, Timeand Condé Nast, WIRED’s parent company, as well as user-generated content platforms like Reddit. (Automattic, which owns WordPress.com and Tumblr, was also said to be in licensing talks this year.) Because the details of those deals weren’t disclosed, it’s unclear how huge their individual corps are.
When WIRED asked about the size of the datasets it would be considering for commercial licensing, OpenAI declined to share that information. However, spokeswoman Kayla Wood emphasized that the company’s partnerships with publishers “are focused on serving their content in our products and driving traffic to them.”