Scraped and Fighting Back: Justine Roberts CBE, Mumsnet CEO, on Fairness and AI’s Potential

Parenting Meets AI: Mumsnet‘s High-Stakes Battle For Survival

Mumsnet might not be the first name that springs to mind when you think of artificial intelligence. Yet the popular parenting platform finds itself at the centre of a growing, industry-defining conflict, balancing how to harness the powerful technology for innovation while staring down what CEO and founder Justine Roberts CBE described to FutureWeek as an existential threat to publishers.

On one front, Roberts said Mumsnet is fighting for its very survival against the AI startups and tech titans scraping its treasure trove of data without permission and it is preparing for a landmark legal battle against OpenAI. On another, the platform is embracing AI to unlock powerful insights from its unique content.

Mumsnet’s complicated relationship with AI reflects the wider tensions in publishing, where the stakes for Roberts—and the industry—are high. “We absolutely see the potential benefits not just to our own business but to humanity,” she said. But Roberts warned AI also has potential for great harm and said the benefits don’t “mean it should be allowed to march forward without any kind of restraint or thought or restriction by governments and lawmakers.”

Mumsnet vs OpenAI

Mumsnet first realised it may have an AI problem in April, when it discovered OpenAI, the Microsoft-backed startup behind ChatGPT, had been scraping its site for data. Large language models like GPT, the backbone of AI tools such as chatbots and content generators, rely on vast amounts of data. Much of it is pulled from publicly available sources like Mumsnet, raising questions about consent, copyright, and compensation for the creators and owners of that content.

The forum’s archives are a particularly valuable resource, Roberts said, with around 6 billion words of conversational data. These words feature real, unfiltered discussions on an array of topics including parenting, relationships, careers and products. Unusually for the internet, it is dominated by women. Roberts said this allows women to be “very honest about their experience” and makes the data an incredibly valuable resource.

Roberts said the team took steps to stop further scraping and reached out to OpenAI to discuss licensing the data. She said the company went away and had a “proper look” at Mumsnet’s data. First impressions were promising, Roberts said, but OpenAI returned six weeks later after concluding the dataset was too small and too open for its needs.

Roberts, who said an NDA prevented her from discussing details of the negotiations, doesn’t seem to buy it. “I think has been articulated a number of times now that they don’t believe they should have to pay for freely available information on the internet,” she said. “They could easily afford to come to an agreement,” Roberts added, noting that OpenAI was valued at around $157 billion in October after its latest funding round.

Mumsnet is now weighing its legal options, and is prepared for a fight if necessary. While there are plenty of active copyright lawsuits against AI companies, Roberts said Mumsnet is unusual in that its servers are based in the UK. This makes the fight a potential test case for British copyright laws that could set a global precedent, influencing how governments and courts regulate AI’s use of publicly available data.

Taking on Big Tech is expensive. Roberts explained that the tech industry’s playbook is to make legal cases as lengthy and expensive as possible, but that “justice shouldn’t only be available to those with deep pockets.”

In the long run, Roberts said Big Tech could damage themselves by scraping data without compensation. “There’s a fairness argument here. If AI companies continue to extract everyone’s data from the web and build products that can replace the websites from which they’re getting this content, without offering financial recompense, the contracts will all dry up. What will they have left to train and verify their models?”

Building MumsGPT

Ironically, as Mumsnet prepares for a legal showdown over AI’s use of its data, it’s also fully embracing the very technology Roberts has branded an existential threat to publishers.

While it would be an understandable position, Roberts is the polar opposite of anti-AI. Before Mumsnet had even realised its data was being hoovered up by tech startups, it was busy building its own AI tool to draw insight from the forum’s wealth of material. Enter MumsGPT.

The tool’s name is now laden with irony in light of the company’s legal battle. It has not gone unnoticed to Roberts. “Obviously we are using the paid version of these things and because it’s a tool we are happy to pay for it in the building of our own stuff.” The company has since switched to using Google’s Gemini tools, Roberts told FutureWeek.

The plan is to glean insights from Mumsnet’s array of conversational content. The anonymised, female-dominated data could be a goldmine for companies trying to understand household decision-makers, most of whom, Roberts explained, are women. It’s also useful internally, Roberts noted, and can be used to help produce thought leadership pieces and reports.

It’s hard to say whether MumsGPT will be deployed more widely at this stage, Roberts said. “It’s not clear whether it will end up being an external product or (stay) an internal tool.”

“We’d love it to be something that was useful to brands and they were prepared to pay for.”

Balancing Innovation, Ethics and Survival

Mumsnet’s journey with AI reflects the broader challenges publishers face in the era of generative AI. On one hand, the technology offers untapped avenues for innovation, opportunity and growth. On the other, it raises thorny questions about ethics, fairness, and control.

For Roberts and her team, the stakes are deeply personal, too. Mumsnet was built to amplify women’s voices. The company’s mission—and its dataset—is a rare and valuable counterbalance to the male bias characterising the field and the outputs of some models. “Having more women’s voices reflected in the models can only be a good thing,” she said, helping AI better reflect diverse perspectives. But including this without compensation or permission risks eroding trust in platforms like Mumsnet and undermining the very diversity that makes such datasets valuable.

AI is going to be as transformative as the internet, Roberts believes—potentially even more so. But she warns that without thoughtful regulation, ethical safeguards and fair systems of compensation, its harm could rival its benefits. “All of us will be spending the next 20 years getting to grips with it,” she said. The challenge ahead, she emphasised, will be striking the right balance—unlocking the opportunities of AI while ensuring a fair and sustainable future.

Subscribe to our newsletter for updates

Join thousands of media and marketing professionals by signing up for our newsletter.

"*" indicates required fields

Share

Related Posts

Popular Articles

Featured Posts

Menu