That is certainly the question du jour in our industry and this is one soup I do not want to eat. My book, Styled, was included in the now-infamous Books3 data set – It was one of the 183,000 books that were used (stolen?) by multi-billion dollar businesses like Meta and OpenAI to train their AI models. The entirety of Styled was ingested so these language models could, you know, learn about design and styling.
Last week I went to New York with Raptive, our programmatic ad network, and partnered with 33 other content creators to launch a campaign to regulate AI online. You see, these major tech companies have built tools that have scraped (and memorized) the contents of the entire internet. These large language models, as they’re called, have ingested every publicly available photo, every paragraph, every recipe, every instruction, and every DIY; now, with the simplest question typed in a search bar, the AI model will regurgitate a generic response, without crediting the sources of its information. This isn’t just a problem for those of us with internet jobs, though – if you’ve ever shared anything online publicly (a photo, a video, a comment to a friend), it has already been used to train at least one AI system.
Obviously, AI is here and certainly playing a victim to it isn’t the answer – we aren’t crying and obviously continuing to just work harder to innovate, diversify, and keep long-time readers coming back. So last week, in NYC, I joined a panel with Tieghan of Half Baked Harvest, Kaitlin of Woks of Life, and Kevin of Country Rebel – representing all of the content creators out there whose businesses are being directly impacted by these large language models.
You see, AI is capturing new readers that historically have stumbled upon our human-led sites by typing in “how do you make hummus” or “what size for living room rug.” Now most of us are getting ahead of this already (we’ve been trying to create non-AI-able content for the last two years which I wrote about here), and it’s our job to create such compelling content to keep readers coming back HERE every day – “direct traffic” is our north star.
But what our digital media industry is seeing, as a whole, is a newfound inability to turn searchers into readers. Said another way: let’s pretend that you’re just getting into design. Maybe you’re decorating your house for the first time, and you Google a question like “how high should I hang my curtains?” In the past, a reader might click on a link to our site – and maybe they’d like it, and they’d sign up for the email list, and we’d earn ourselves another regular reader! But now, Google’s Gemini model is serving up our words in a very generic form – the searcher is never even given the option to check out our site.
Ironically, EHD’s search traffic is up in 2024 – our team has been working really hard to make sure that readers can still find us in a sea of generated answers, for which I’m so grateful. But that’s not the case for many other online digital media companies – there are major drops in traffic to more evergreen (year-round, year after year) posts because of AI. We can all see the writing on the wall, so I’m honored to be part of the group working to find solutions on behalf of thousands of small web businesses out there.
Will we be fine? Well, ask print magazines (gone due to blogs – mine included), local news and newspapers (gone because of podcasts and digital media sites), and the early 2010s design blogs (killed by social media and Pinterest). The “information industry” always changes, staying ahead of it has been a very exciting challenge over the past 15 years and it’s just part of social evolution. Being a large player in this industry for this long is remarkable on its own. And listen, I’ll write this blog until I die because I love the hell out of being right here.
As the quantitative futurist Amy Webb recently said on Brene Brown’s podcast – navigating AI is like driving on ice, slamming on the breaks will kill you, you have to steer into the spin. Control the chaos and remain calm. The ice is cold and hard, but it’s here, and unless you want to get out of your car and walk you better learn how to drive on top of it to get where you want to go. I feel confident that Raptive – the tenth-largest digital media company in the world, with an online reach that’s larger than that of Hearst, Twitter, Fox, Reddit, etc. – is well-equipped to negotiate an agreement with AI companies that preserve the richness of humanity on the internet.
What happens if they don’t?
Well, what I warned last week (and what I firmly believe) is that the following three things will happen in this order:
- Food, parenting, decor, lifestyle, and fashion blogs will lose traffic and revenue. Creators and our teams will have to find new revenue ASAP. If we can’t, our businesses will close. All doable, but sad for true followers.
- The internet will become wildly uninteresting. AI will essentially be its own parasite, sucking the entertainment and fun out of digital discovery.
- As lifestyle websites powered by humans disappear, so will the quality “answers” that the AI has been scraping. Instead, it’s likely to serve up inaccurate and certainly un-nuanced answers to your questions. (This is already happening in a small way – like when Google’s Gemini suggested that searchers eat glue and rocks – but researchers know that it has the potential to get much worse, too. When AI models run out of information and begin ingesting AI-generated answers as a new data source, the models break. The entire internet, as we know it, would be rendered unnavigatable and unusable.
- Bored by the now generic internet, people will simply pick up their phones and go to TikTok – the platform du jour that has been engineered to be quite an incredible search engine (Instagram, less so). Do I think that Google will disappear? Nope. Those folks are genius and they’ll figure it out, as long as they’re not eaten by a skilled AI model (like Claude, by Anthropic) or a rival search engine (like Perplexity, which functions exactly in the way you probably imagined Ask Jeeves to work about 25 years ago).
What do I think should happen? Glad you asked.
The legal landscape here is fraught, which makes this tricky: tons of AI companies have been sued for copyright infringement and it’s unclear how those court cases will shake out. But I do think that the current business leaders in AI – Google, Meta, OpenAI, Anthropic, Mistral, etc. – have an opportunity here to partner with human creators who have been and hope to continue to create real original content. Serve human responses higher in the algorithm, above the AI-generated answer. Pay us directly for our work, or license it, instead of stealing. Work directly with digital media experts, like the team at Raptive, to identify creators who are adding value to the internet ecosystem. To be fair, there is also some hope here! Caitlin has had some great conversations with startups like ProRata.ai, who are working on attribution technology – basically ensuring that creators are credited (and compensated) when their information is used to power profits for a multi-billion dollar AI business.
I don’t want to sound all doom and gloom. AI has some incredible applications. It can sort through extraordinary amounts of information and it can make connections that seem invisible to humans. It can (and will) radically transform medical diagnosis and treatment. It can empower human learning, creativity, and exploration.
But in the digital media space, it falls short. So I’m advocating for the creators who add richness to our lives – creators who bring us killer dinner party recipes, creators who help us feel more comfortable in our bodies, creators who take us behind the curtain and make us feel less alone. I’m advocating for my fellow authors, whose books were stolen and whose words now shape every answer spit out by a generative AI model (including mine). And I’m advocating for you, too: AI companies have trained models on information stolen from all of us, without offering us an option to opt out. I think we all deserve better. Let’s keep the humanity in the internet – we all know that’s more interesting to read about, anyway ? What’s your take?
(If you’d also like to help keep the internet interesting, you can sign Raptive’s open letter on AI here. Every signature helps as they continue to advocate for a human-first internet.)
Also, here is a list of all the creators that attended the conference! Please check them out, enjoy their real content, and give them a follow if you like their stuff:)
*Photos via Raptive