What We're Telling Publishers About The AI Content Apocalypse

Or, The Return of IP Lawyers

G/O Media, the private equity-backed holder of the former Gawker Media properties (as well as Quartz, The Onion, and others) apparently told its staff yesterday that it was selling Lifehacker.com to Ziff-Davis.

Lifehacker, founded in 2005 by Gina Trapani (as an employee of Gawker Media), has been since its inception a shining example of service blogging: helpful, lucid, and perfectly suited to SEO. “How do I…” searches funneled millions of readers to Lifehacker’s content from Google for almost two decades; display and affiliate advertising generated millions of dollars in revenue from that traffic.

It is the kind of site that every media group wished it had (alongside a gadget blog or, later, a gadget shopping guide). Relatively inexpensive to operate, inoffensive, rarely provoking legal suits or public controversy. Money in, more money out. The kind of site, in fact, that kept Gawker Media afloat even when the gossip blogs operated at a loss.¹

While the price Ziff paid hasn’t been announced, it’s safe to say that the sale signals at best troubled financial forecasts for G/O Media, and likely the process of dismantling the network and selling the parts.

If you think you can sell the whole network at once, you don’t sell off one of its most reliable revenue generators first.²

---

For the last few months, my consultancy³ has been approached by several “blog networks” or—on the other side of the deal—private equity groups considering purchasing them.

Some of the networks have been trying to figure out new revenue streams since post-pandemic search traffic and ad revenue has cratered. Or they’ve been looking for connections to private equity or publishers who might be interested in an acquisition.

Few of the existing networks are interested in merging. Even the networks that established themselves in the last few years buying up distressed brands—Bustle, Recurrent, Penske—aren’t picking up new properties in 2023. (Many of them have too many brands that they’ve bought and never re-monetized already.)

The revenue sources of online media for the last decade—subscription, display, affiliate—were never great to begin with. They’ve only ever been the scraps left over from big advertising whales that Google, Facebook, and Amazon have already consumed. And this year, the carcasses floating to the sea floor are arriving picked clean.

Add that many of the “newer” online media networks—Vox, Vice, Buzzfeed—took tens and hundreds of millions of dollars in venture capital to launch and then never sold a decade ago when they had the opportunity—and are now seen as failed businesses (by VC standards)—and online media as we’ve known it since the Web 2.0 days is not looking too healthy.

And that’s before chatGPT entered the picture.

---

The innovation of commercial blogging was largely made possible by three things:

1. A workforce of untrained, non-union writers who would work for wages far below industry standards.

2. Speedy publishing, made possible by foregoing a traditional fact-checking process or (often but not always) original reporting.

3. Cheap publishing software.

The tone and perspective of outsider voices was valuable, of course, but it wasn’t an essential pillar of the business model.

As soon as Google formalized their claim over online advertising with the acquisition of DoubleClick, the system was established: make content that indexed well in response to search traffic. The revenue and retention strategies of online publishers have evolved in response to stimulus from Google and later Facebook and Amazon, but the core loop never really has. Get your content in front of as many people as possible, on any platform or channel possible, and monetize the clicks.

It was clear about a decade ago that the chief product of online publishing—words in a blog—didn’t reach or convert as many people as video or platforms (and their apps). So Instagram sold for a billion (a steal); Facebook became nation-state-sized; small creators built better businesses on captured channels like YouTube or TikTok with video than writing ever transacted online.

Many creators were just repackaging old content in new formats, but it worked, and it happened that the skillset to make clicky video rarely overlapped with the ability to write a good headline.

---

Here’s what my partner and I have been advising publishers asking for our outlook on the future of “traditional” online publishing in the wake of chatGPT and other AI/ML-powered products and interfaces:

1. You need to figure out what novel content you own and what it costs to produce it. And if it is original reporting, invest in that.

2. You need to hire lawyers, individually or in coalition with other publishers.

3. If your model is predicated on tonnage—being one of the hundreds of possible places Google sends a searcher asking a generic question by creating thousands of discrete posts that give an answer—you may not make it.

The reason that Microsoft and OpenAI’s syndicate is negatively impacting the perception and stock price of Google is that it is now possible to see a future where “content” doesn’t leave the “search engine.”

Right now, in multiple apps and websites that I use daily, I can type something like “/ai” or open up a prompt and ask an AI to generate copy for me. The results are usually legible, occasionally clever, and sometimes even accurate.

It’s good enough.

And let’s not pretend that most of the content created by humans in the online publishing ecosystem is good. The inexpensive nature of the format; the failure of (or disinterest by) Google to police low-quality content, which largely serves as window dressing for its own paid links anyway; the machine generation that was already in place by spammy link farms; the fact that most people now search for “The Product I Want Reddit” to try to find at least some genuine consumer feedback instead of being shunted to guide or faux review sites—all of these are systemic weaknesses in the ecosystem already.

People aren’t stupid. They’re busy. If typing a phrase into Bing gives them 1,000 words and some images that answer their question, they’re not going to miss going to your site, even if it’s a 100-year-old brand.

Which is why we recommend lawyers and perhaps lobbyists.

The one thing that AI-generated content is bad at (for now) is chain of custody for its source material, its corpuses, and copyright.⁴

There are a few early copyright cases that are promising for content creators and publishers. But the lesson from the last two decades wasn’t that copyright alone will protect your business model, but that current transformative rights and case law—a YouTuber using your reporting as the basis for their video; another blogger summarizing your original reporting; a robot using terabytes of your magazine’s copy to build a predictive language model, perhaps—also covers creators and publishers who “remix” your original content.

(Ironically, many of the early blog publishers, myself included, bent the strictures of copyright and lobbied to increase the scope of fair use in the early 2000s.)

There is a presumption from name-brand publishers—major news outlets, for instance—that their work will withstand the now effectively limitless tonnage that can come from AI, not just because of their strong brands and reputation, or their lawyers, but because they traffic in verifiable facts. That will be as true for the largest journalism brands as it was untrue for most of the smaller regional newspapers and conglomerates who have already gone out of business.

But the near-term opportunity for the publishers who can afford it (individually or as part of some sort of industry-level consortium) will be to hammer out pricing models for the information that powers the AI products (that will only increase in usage) with the tech giants who are large enough to have money to share and can’t risk falling afoul of future law or lawsuits.

That may be the most practical hope at this point. Online media of the last decade isn’t a great business, especially with highly paid writers, offices, and support staff (as lovely and fair as that all is) and is largely kept alive by downstream revenue from tech giants. But those giants are about to go to war again.

As I said to one media C.E.O., “I would suggest flying to Redmond and trying to be the first publisher to strike a licensing deal with Bing.” Will it work out for everyone? Absolutely not. But the first shops to pivot to video got several large checks from Facebook for a couple of years that, spent wisely, helped them stay alive a little longer.⁵

1. They always operated at a loss lol

2. A counter-argument: Perhaps G/O née Great Hill purchased the network for such a fire sale price from Univision that selling even one of its marquee properties is enough to justify the whole purchase. Plausible, but short-sighted.

3. No, we haven’t announced our new firm; No, we are not offering management strategy for publishers, but we are connected in the space so we get a few calls.

4. Can you imagine the Curator’s Code launching in 2023 lol

5. This is what is commonly known as a “lifehack.”