Editorial: Generated AI’s use in farm and food content

The growth of generative AI content over the past two years has been undeniably substantial — it can be found in all corners of social media and even on some major publishing platforms, with Sports Illustrated notoriously being one of the first and biggest offenders.

And yeah, we said “offenders.” Without some rules and transparency guidelines in place, AI content can be pretty dicey, with some onlookers even referring to its use as “outright dishonest.” It can also be very easy to confuse an audience and to use AI for underhanded purposes, whether that means to deceitfully tug on someone’s heartstrings or to make someone (or something) look bad.

A couple of months back, we probably all remember this viral image of a little girl clutching a puppy. It garnered a lot of sympathy in the wake of Hurricane Helene hitting the East Coast last fall — the image plastered across social media, particularly Facebook. But, that image was AI, and many people didn’t realize it immediately.

It was a con, pure and simple:

And if you think that those of us in agriculture are immune, then you’d be wrong.

Artificial intelligence is justifiably being used in various sectors of our industry (like here and here), yet on the content side, we’re seeing more and more pressure to use generated AI information. Conference workshops are being built around generated AI content, and there is a growing need for media outlets to play into the “AI Overview” that Google searches dish out — without ever having to click on an article or send a pageview to the site that originated the content.

It would be too easy if it ended there, though. Unfortunately, over the past month, we’ve seen social media erupt in new and creative ways to play on emotions related to food and farming, often with the farmer on the losing end of the poster’s message.

And just like with the girl and her puppy, these images are not always obvious to someone who isn’t approaching it with skepticism. Take this one for example:

Gobs of likes, comments, and shares for something completely made up and done so to manufacture insult and outrage at being slighted as a farmworker. AI even makes sure that every one of the young women pictured is pretty because that makes the “injustice” depicted here all the more poignant. (At least it looks like AI got their hands mostly correct, which is where you can usually tell if something is AI-generated.)

Some of the ways AI is used is to create ridiculous images, no matter whether you agree with the message or not.

It’s painful, but it’s everywhere these days. And the worst part is — I don’t follow any of the pages where these images showed up, but Facebook served them to me as recommendations based on other (loosely related) pages I do follow. So if I follow an agricultural page, of course I want to see everything about Monsanto, right? 

Wrong. 

I don’t follow these types of sh!tposting sites because they are intentionally misleading and create outrage either for the sake of clicks or so that they can sell some alternative “natural wellness” product and make a few bucks.

Facebook isn’t doing us any favors a lot of the time by suggesting these posts, but at least on some of the most egregiously trashy posts (like the Monsanto one above), you’ll see a small “AI info” label on there for the sake of transparency. It’s something, but Facebook could — and should — go further to address this kind of sensationalism.

Yet it’s more likely that it will all get worse, not better.

I understand that generated AI isn’t going anywhere, and more notably, it is going to be more and more the norm in the coming years (will ag writers still have a job in five or 10 years?). It’s a cycle that feeds into itself because generated AI aggregates existing content on the internet to “inform” the content it creates. That’s why, for example, it’s so hard for generated AI to depict someone writing with their left hand — most people are right-handed, and AI harvests known images, usually righties, to artificially create the photo you ask it to make.

The AGDAILY staff hasn’t yet embraced generated AI content on the site, and there are a couple of reasons. Primarily, it’s about trust and authenticity. Not to mention that, because AI harvests existing online information, we would risk pulling from a search database of historically polluted information from groups such as the Environmental Working Group or the Humane Society of the U.S. And we’re sure that any farmer or rancher BS meter would be on high alert with that kind of content being presented.

We ask that everyone stay vigilant about the kind of content they come across and decide for themselves who to trust.

Ryan Tipps is the founder and managing editor of AGDAILY. He has covered farming since 2011, and his writing has been honored by state- and national-level agricultural organizations.