Meta's AI scraping debacle should be a warning for news outlets
Meta might as well be Taylor Swift for all the eras it has gone through over its existence. I hate to go all Grandpa Simpson, but back in my day, a breaking news badge on Facebook could send one post into the stratosphere. For local news and niche marketers, it served as an important part of an alert strategy. Then Meta restricted the number of stories that could use the badge per day before ditching the badge altogether.
Now Meta may be entering a new dark era, one that should have publishers and any content creator rethinking their future with the platform. Earlier this month, Nieman Labs reported that Meta's updated its commercial terms of use, which covers any business accounts on its popular platforms, to allow the company to "use web scrapers to crawl and collect content from websites 'associated with' business accounts on its platforms." The language specifically gave Meta the right to "train" its AI models on the website content. The changes were set to go into effect on Oct. 30, and there was no way to opt out.
"As written, the new terms could be used to argue that Meta can scrape content from news sites and use that content to train the company’s AI products, according to three legal experts who reviewed the terms at Nieman Lab’s request," an Oct. 15 report said.
So under this agreement, anything my team shared on Facebook or Instagram could be scraped by Meta's agents and used to train its AI models. For what purpose do those AI models serve? Well, right now maybe nothing, but I could already imagine a future in which Meta built its own news platform and cut out publishers altogether. Nearly every outlet has at least one Facebook, Instagram, Threads or WhatsApp account.
Immediately, I started running through scenarios of how to get my newsroom off the platform, but whatever I could try, it couldn't be done quickly enough to offset that stinging traffic loss.
For hyperlocal news outlets like mine and other niche market creators, AI represents a double-edged sword. In search, for example, the AI overviews haven't completed kneecapped our strategy. Sure, there's been a drop in traffic, but an increase in subscriptions from search. A search audience is a more driven, purposeful one, and if they find the news they seek, they'll probably pay for it. With so few outlets left, the hyperlocal ones still standing rank highest and see their links featured in the overviews, right where people click them.
Of course, on the other hand, the industry remains extremely vulnerable to AI, as Meta demonstrates, and it's not hard to imagine Meta using local news to train its AI models and build its own news platform.
Meta has since walked back its language, and the good people at Nieman have been holding their virtual feet to the fire. But publishers and creators shouldn't get comfortable. The whole incident highlights the gamble of putting all your eggs in one audience distribution basket, especially one platform that owns many. Gmail may be the dominant email provider, but if it fell tomorrow, I could still dust off my msn.com account.
When a platform changes its terms, you're stuck with them unless you can pivot fast enough. I know my team couldn't. We might lose (a lot, admittedly) traffic now, but we might lose everything in another two to five years if we stick around and help train that which could destroy us.
Comments