February 21, 2023

Does AI Put Evergreen Content at Risk?

With any new technological advancement, an individual can find themselves on one side of the impending conflict. You are either clinging to a past that is already long gone or you are the vanguard of a charge forward into an unknown future. And it is no different with all the recent developments of search engines, ChatGPT, and the impacts on publishers’ businesses.

I, myself, am in conflict about how I feel about this. On one hand, this technology is incredible and the tools being introduced are revolutionary. I already use it to clear out the “blank page” crisis any writer has when trying to determine how to start a piece. I haven’t taken a word from ChatGPT, but the fact that it can help me get grounded is unbelievably helpful. And that’s just the beginning, I’m sure.

On the other hand, it appears to break the social contract between search engines and publishers. Search engines organized all information and sent people seeking details to a publisher’s website. This allowed both parties to monetize their respective platforms. By keeping people within a chat-like interface, but learning from publishers’ content, the platforms get to organize and display the information without having to share the traffic.

Even if a social contract has been broken, though, doesn’t mean that there’s any infringement from a legal perspective. According to a very good piece published on Creative Commons:

In fact, the Supreme Court has recognized fair use’s importance in the development of new technologies, first in 1984, in Universal City Studios v. Sony and most recently in 2021 in Google v. Oracle. In Sony, the Court held that the Betamax videocassette recorder should not be sued out of existence even if it could potentially help people violate copyright law. Instead, because it held “substantial, non infringing uses”, the Court believed copyright law should not be used to stop it. Then in Google, the Court held that Google’s use of Google’s 11,500 lines of Java code was fair use, writing that the courts must consider fair use in the context of technological development.

Altogether, I believe that this type of use for learning purposes, even at scale by AI, constitutes fair use, and that there are ways outside of litigation that can offer authors other ways to control the use of their works in datasets.

It’s a fascinating legal debate and one that I can guarantee the legal system will get wrong one way or the other, in big part because people who are making decisions on cutting edge innovation still can barely send an email.

But with any innovation, there will be winners and losers. Wired published a piece about how media companies are “wary of the Bing chatbot’s media diet.” In it:

When WIRED asked the Bing chatbot about the best dog beds according to The New York Times product review site Wirecutter, which is behind a metered paywall, it quickly reeled off the publication’s top three picks, with brief descriptions for each. “This bed is cozy, durable, easy to wash, and comes in various sizes and colors,” it said of one.

Citations at the end of the bot’s response credited Wirecutter’s reviews but also a series of websites that appeared to use Wirecutter’s name to attract searches and cash in on affiliate links. The Times declined to comment.

It is increasingly becoming clear to me that mass-produced evergreen content is at risk with this technology. Wirecutter writes reviews of products. Forbes does too. Red Ventures. Ziff Davis. Hearst. They all publish tons of evergreen product reviews. The goal? Rank competitively in Google, get traffic, and then hope people buy the products after clicking one of your links.

It’s not a bad business. And there are certainly brands that are better than others (Wirecutter is owned by NYT, after all). The issue is that a product review doesn’t change from day-to-day. Once you’ve made it, it exists forever. And so, a chat-like AI can learn from all of them and when a user asks, “what is the best dog bed” as in the example above, it can provide the information. The AI is curating and giving its audience the information it needs.

The thing is, this is normal behavior and is how most of media operates when it comes to content aggregation. Newsletters do it. Blogs do it. TV news definitely does. In a skit about journalism by in 2016, John Oliver said:

It’s pretty obvious without newspapers around to sight TV news would just be Wolf Blitzer endlessly batting a ball of yarn around. It’s not just news outlets. Stupid shows like ours lean heavily on local papers. In fact, whenever this show is mistakenly called journalism, it is a slap in the face to the actual journalists whose work we rely on. Two years ago, we ran a piece on state lotteries and a not insignificant portion of it was built on the work of Harry Esteve, a reporter at The Oregonian.

The media is a food chain which would fall apart without local newspapers.

That food chain exists in every aspect of media. And so, it should come as no surprise it’s happening again; albeit, at a scale that perhaps most people did not anticipate.

The problem for publishers that are overly reliant on this sort of evergreen content is that it makes visiting their sites less necessary. If ten sites all have the same product reviews, the AI can learn from them all and publish a curated review. Sure, it might include references, but does the reader even care? They got the information they needed.

And this is a legitimate risk for these big publishers. To make matters worse, they’re stuck between a rock and a hard place. The natural response might be, “well, we’ll just block the AI so they can’t learn from our content.” Unfortunately, they can’t do that because blocking the AI means blocking the search engine. And that cuts off their source of traffic.

This is where Skift’s Rafat Ali’s thoughts on AI become very important. In this tweet, he says:

Crappy & average content creators and owners have most to worry about & I give two f**ks for their misery. High-value creators like what I’ve built my companies around, will only increase relevance in a world of glut. Build your brand around quality, not quantity, you’ll thrive, it isn’t just a glib answer, the two companies I have created are proof.

He’s right. Brand is going to play an incredibly big role in this AI-led future. It’s impossible to say what percentage of Wirecutter’s traffic is direct to site because NYT has Wirecutter as a subdirectory, not its own site. However, according to SimilarWeb, 60% of NYT’s desktop traffic is direct. That is entirely a function of brand. People know NYT’s brand so much, they type it in. And I imagine there’s a decent percentage of people who also type in Wirecutter’s brand.

So, while AI might hurt evergreen content, it certainly doesn’t destroy it. It simply changes the equation of where the traffic comes from. More importantly, it forces the conversation from mass producing content for the sake of growing search traffic to actually serving your audience with information they need. If you do this well, you’ll build a brand. Over time, people will begin to think of your brand when they need certain information.

Ironically, there is one group of publishers that I think could thrive here and that’s the news business.

Philip L. Graham, the former President and Publisher of Washington Post is often credited with saying that journalism is “the first rough draft of history.” And in many respects, this makes sense. Before a story is published, there’s no information. And tomorrow, that information might have been supplanted with newer information. And day after day, new information comes out. That gives a news organization an advantage.

That’s not to say, of course, that an AI wouldn’t also be able to learn news. And for many passive consumers of news information, the transition from Google News to Google Bard (its AI) or Bing’s AI would be seamless. However, delivering “get nowhere else” news is a competitive advantage that would help a publisher build a brand.

Much of the internet has been built on the regurgitation of information. AI is no different. And whether we operators like it or not, it’s here. We can’t control what comes next.

Thanks for reading today’s AMO. If you have thoughts, hit reply. Become a premium member to receive even more AMO and for an invite to the AMO Slack channel. Have a great week!