The publishers’ guide to being gaslit by tech platforms (the AI edition)

But Google isn’t alone in rewriting the story. Across platforms, the same playbook is in use: soften the language, reframe metrics, show up to the odd industry event with platitudes. This guide unpacks the tells — how to spot the spin before it becomes the story. 

“Trust us, traffic is actually better now”

Get used to hearing that. Google’s latest blog post was less a defense and more a soft spin: AI Overviews aren’t tanking traffic so much as they’re improving it.

Sure, there’s a case to be made: a click after an AI summary might be more meaningful since they’re likely to be more engaged. But for publishers watching referral traffic crater, it’s a hollow argument. 

The takeaway: when platforms start redefining what success looks like, it’s usually because the old metrics look ugly. Trust your own data, not their narrative. 

“Reasonable licensing terms” means take it or leave it

One of the most reliable red flags in AI licensing announcements isn’t buried in legalese — it’s right there in the language. 

Take this phrase: “reasonable licensing terms”. It shows up often, and it sounds…fine. But what it usually means is: “we’re paying just enough to call this a deal, not enough to match the value we’re extracting”. The uncomfortable truth is that there is no real market rate for AI use of publisher content — only the one set by the platforms. Most publishers don’t have the leverage to define what “reasonable” even means. 

The takeaway: watch for euphemisms. Words like “reasonable” and “fair market value” aren’t neutral — they’re strategic. They reframe power imbalances as partnerships. Learn to decode them. Or if you can afford to, hire an expert to do so like The New York Times, Washington Post and Gannett have. 

But if you are going to take the check, do so knowing its smaller than it looks

We get it. Sometimes you take the check. Just go in clear-eyed: short-term money rarely solves long-term problems in media. These deals might smooth things over for a minute but they don’t necessarily fix the underlying cracks. If anything, they usually leave more questions than answers.

As one one commercial exec put it on the condition of anonymity: “Are these deals intended to be a gesture to the news community, which is going to aggravate hundreds of sites and media companies and satisfy 20?”

Which is why some publishers are playing it slow, waiting to see whether a real market around these deals emerges. Optimists think platforms will come around. Cynics are betting on lawsuits or regulation to force their hand.

The takeaway: Only a handful of publishers will get real money — News Corp is getting $50 million a year from OpenAI, for instance. The rest will likely get gestures. And while those checks might help now, they won’t rebuild what AI is eroding. 

“Your Content Is Powering the Future (But you’re not invited to shape it)”

Publishers have long feared that their content would become raw material for someone else’s roadmap. The platform has done little to prove them wrong. Every few years, a new product cycle rolls around, the pattern repeats: media fuels the machine but never gets to steer it. The AI boom is the latest iteration of that dysfunctional flywheel. 

Case in point: OpenAI’s Nick Turley recently said that the company’s partnerships are meant to “shape the future of content”. Days before, OpenAI skipped an IAB Tech Lab event where 80 media execs gathered to discuss exactly that. The optics aren’t subtle. 

The takeaway: Platforms will pay publishers for content, but as for giving them control? That stays in their hands. 

Showing up is not the same as showing transparency 

Platforms are getting better at optics. They’ll do the industry events, they’ll join the panels, they might even sponsor the conference. But don’t confuse presence for openness. 

Look at Meta at last month’s IAB Tech Lab event last month. The exec who presented there – Shelley Venus, Meta’s senior director of global media strategy – was polished, on-message and said all the right things for the room, according to two execs who were there. But, as always, the proof isn’t in the pudding. 

“[Venus] was clear that the new leadership knows that AI runs on good content, and that has moved Meta to engage more [with publishers],” one of those execs told Digiday, under the condition of anonymity. “TBD of course on what that looks like.”

“We remain committed to supporting quality journalism”.

This is the closing line of countless blog posts, press releases and on-stage remarks. But like any good PR line, it’s so vague it’s functionally meaningless. 

Supporting journalism can mean a licensing deal with a handful of publishers, or a PR grant to fund newsroom innovation. But it can also coexist with algorithm shifts that gut referral traffic and AI products that summarize articles without sending users to them. Commitment, in this context, doesn’t always come with accountability. 

The takeaway: Look at the incentives, not the statements. If a platform’s business model thrives while yours erodes their “support” isn’t a strategy — its a slogan.