Google Paying Publishers to Use AI Is Bad, but Not for the Reasons You Think

Credit: DALL-E, assisted by Photoshop Generative Fill

The news last week that Google is making deals with several smaller newsrooms to get them to use the company’s generative AI tools to publish stories was met with near-universal skepticism from the media industry.

This is certainly understandable. For starters, Google never announced it was doing this, so when Adweek broke the story, there was almost scandalous nature to it. If Google thought it could just do this quietly, that was a pretty naive stance. Keeping secret a deal with an industry whose entire incentive system is to reveal confidential information… well, that’s a hell of a blind spot.

Add to that the bitter resentment the media industry has toward platforms like Google, which have largely architected a system where they feast on nearly the entirety of the digital advertising market while leaving mere scraps for publishers. The news media got the short end of the stick the last time there was a technological upheaval, so it’s only natural to be wary of it happening again.

Still, looking at the nuts and bolts of what Google is doing, there’s a strong case to be made that its efforts to empower newsrooms with AI are a good thing. The focus is apparently smaller, local newsrooms, which have been largely decimated in the digital age. Most local publications have next to no editorial staff, so anything that amplifies — or even multiplies — the productivity of those few people is direly needed.

There’s also the basic philosophical truth that journalists and journalism need to accept that generative AI is part of the media ecosystem now, and that all news organizations will need to at least account for it, if not actively use it. There are many use cases besides pure content generation (and using it to produce content at scale is certainly a bad idea), and Google’s tools seem to be mostly focused on news gathering, doing things like discovering and summarizing important takeaways from long public documents (think: school board and town council meetings) — a relatively benign exercise that still requires human oversight to create content.

Subscribe now

Google’s Bad Medicine

However, publishers are right to be wary of Google’s gift, which is rumored to be worth tens of thousands of dollars for each newsroom they “empower” with these tools. Google certainly isn’t motivated by altruism or some newfound respect for journalism; it obviously wants to lock customers into its Gemini models, its OpenAI-rivaling LLM tech that which the company has fallen all over itself trying to boost.

If last fall’s boardroom drama at OpenAI taught us anything, it’s that locking yourself into a single model is a bad long-term AI strategy. All models are trained and tuned differently, varying widely not just in terms of performance, but also safety features and speed. A workflow with a variety of use cases will likely want to use a variety of models.

In addition, even if you believe that Gemini works well for all newsroom use cases (a big if), the fact is that the incentives of platforms will change. We saw this in the social media era, when Facebook infamously set off a mass “pivot to video” as it promoted short and live videos in its news feed, enabling publishers to reach tens of millions of viewers almost overnight. Of course, Facebook eventually decided it didn’t need the news media, killed all its publisher deals, downranked actual news in the news feed, and shrugged at the industry wreckage that resulted.

A similar outcome here feels all but inevitable. As much as it’s certainly convenient for publishers to accept Google’s money, at some point that money will run out, leaving them locked into tools that may no longer be supported.

The better choice: the news industry needs to chart its own path with regard to AI. And it has been: The AP’s Local News AI program, to name just one example, helps local newsrooms build AI-driven tools. And there are several organizations staffed by journalists (including us!) helping empower journalists and media companies with AI. if you’re looking to incorporate AI into your newroom, feel free to book a free consult.

I get that times are tough, and newsrooms need money. But accepting Google’s gift is the first step toward losing yet again what the media so desperately needs: control over its own destiny.

Ready to start using AI like a pro?


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.