Imagine our surprise, as we were putting the finishing touches on this post when we stumbled upon the news that Merriam-Webster named slop its Word of the Year. The timing felt almost too perfect.
The dictionary defines slop, in part, as low-quality or carelessly produced content. In other words, exactly the outcome many people worry about as AI accelerates how information is created, packaged and pushed into the world.
Every day we see breathless commentary about how AI is changing the way we live, work and engage with the world. Alongside the hype is a fear that machines will replace humans everywhere except the most high-touch professions like surgeons and plumbers.
As moms of teenagers, we think about this a lot as we imagine their futures. And while AI will absolutely reshape work, it’s still a long way from replacing human judgment, context and accountability in many fields.
In fact, we’re seeing more evidence of the opposite. AI works best when humans stay firmly in the loop. Lately, that’s been most obvious to us as heavy news consumers, in one very specific place: photo caption slop.
Exhibit A: When AI fills space but adds no value
A December 15 Forbes article headlined “Are You on Track? Average Salary by Age and Net Worth” included a stock image captioned:
“A person reviewing their finances on a laptop, comparing their salary and net worth to age-based benchmarks to plan for the future.”
That’s not a caption. It’s a speculative internal monologue layered onto a stock photo. Motivation, intent and outcome are all invented.
Exhibit B: Editorializing emotions
A December 6 People magazine article titled “This U.S. Airport Was Named the Most Stressful in the Entire World, Study Says” featured a stock image with the caption:
“A woman stressed out inside an airport.”
This directly contradicts long-standing journalism guidance not to assign emotions or assumptions to people in images. Readers don’t need help deciding what someone is feeling. They need facts.
Exhibit C: Context matters. A lot.
From a December 15 CBS News article on potential cannabis rescheduling, a stock photo ran with the caption:
“Man rolling marijuana cannabis joint in coffee shop with CBD weed buds in glass jars.”
If you know anything about hemp, cannabis regulation, or what’s legal in U.S. coffee shops, the problems jump out immediately. The image might be passable. The caption is not.
This is how slop sneaks in
None of these examples are catastrophic on their own. They’re small. They’re fast. They’re easy to miss. And that’s exactly why they matter.
AI-generated content that isn’t reviewed, edited and grounded in real-world context doesn’t just sound lazy. It erodes trust. Quietly.
To be fair, newsrooms and communications teams are under real pressure. Resources are shrinking. Layoffs are real. AI has been sold as a productivity miracle while many professionals are being asked to do more with less, faster than ever.
That’s precisely why guardrails matter now more than ever.
How to avoid the slop trap
If you’re using AI anywhere in your content workflow, especially in journalism or public-facing communications, a few simple practices go a long way:
- Never let AI be the final pass. Treat it like a first draft assistant, not an editor.
- Be ruthless about captions and headlines. If a line adds speculation, emotion or invented context, cut it.
- Default to factual description. Who, what, where. Skip the why unless you can verify it.
- Pressure-test for domain knowledge. If the topic involves regulation, health, finance or law, assume AI will miss nuance and its output requires validation.
- Slow down the last mile. Most damage happens at the finish line, not the brainstorm.
AI without humans running interference produces unfiltered slop. Maybe that changes someday. For now, human judgment is still doing the heavy lifting.
And honestly, that’s reassuring.
If you’re thinking about how to use AI responsibly without compromising credibility, trust or standards, let’s have that conversation.