The comment that turned it around
In the middle of the crossfire, a social media manager chimed in. Roughly: "Let's recap. Someone wants to help a longtime friend find a job, writes her a recommendation, and the comments start debating whether her face in the photo is authentic enough. A few years ago she'd have gone to a photographer, he'd have retouched it, and nobody would have asked whether it's still her. Today using AI is becoming a question of professional integrity." And the punchline: "She's looking for a job. Not a jury for the most authentic profile picture contest."
I couldn't have said it better. Period. I just replied "Finally, someone normal!" and went for a coffee.
She wasn't alone. Another commenter asked the critics how AI editing differs from Photoshop, a phone app, a filter or a photographer's work, as long as it doesn't distort reality. Then she added: "If AI isn't supposed to make work easier, what is it for?" She called the whole thing hypocritical, at a time when the internet is full of AI slop that this photo has nothing to do with.
Now the interesting part. When I counted the comments, only about 5% were about the photo. Most people simply helped: they shared her profile and connected with her. On LinkedIn that ratio is usually much worse. But those few loud voices got the most attention, because outrage always gets priority in the algorithm. Five percent is enough to make it look like the whole internet is debating a photo.
I call it the crow effect. When someone has nothing to say on the matter, they join the cawing. The nonsense gets lost in the crowd, and from the outside it looks like an opinion.
A year ago: wow. Today: ugh
I'm not alone. The biggest brands in the world are riding the same wave, just with a few more zeros on the production budget.
- McDonald's in the Netherlands pulled a fully AI-generated Christmas ad in December 2025. It lasted three days. People called it "creepy" and "inauthentic"; the company first switched off YouTube comments, then took the ad down everywhere.
- Coca-Cola took heat for its AI-made holiday ads in both 2024 and 2025. A brand that built Christmas on nostalgia was accused of "sacrificing nostalgia for algorithms."
- Meta started tagging photos on Instagram and Facebook with a "Made with AI" label in 2024. Except the label also showed up on photos where a photographer had just removed a speck of dust with an AI tool. After a wave of complaints from photographers, Meta renamed it "AI info."
That last one is exactly my story. The line between "I edited a photo" and "I made a photo" has blurred. And when people don't know where the line is, they throw rocks at everything, just in case.
It's not about technology. It's about trust
Why so much emotion over a hairstyle and a background? In 2024, researchers at Washington State University ran experiments with more than a thousand American adults. They showed people identical products, like a smart TV, and the only difference was the words "artificial intelligence" in one description. The result? People were less likely to buy the product described with AI. The reason, according to the authors: lower emotional trust. And the effect was strongest for expensive, high-risk purchases, where people are afraid of getting burned.
That's the key to the whole thing. People don't see technology in an AI photo. They see two things: deception ("you're showing me something that isn't true") and laziness ("you didn't even think I was worth a photographer"). Both kill trust. And trust is exactly what marketing sells. You with me?
By the way, my line "Btw, AI didn't write this post. I did." was practically an invitation in this context. I told people to think about AI. So they did. :-D
Déjà vu: we've been here before
If this reminds you of the early 2000s, you're right. For over a decade, people complained about "photoshopped" ads: airbrushed models, stretched legs, impossible bodies. Since October 1, 2017, France has even required commercial photos with altered model bodies to carry a "photographie retouchée" (retouched photo) label, with fines of up to €37,500, more for companies.
And today? Nobody cares about retouching. People only care where it lies: when it promises a body, skin or result that doesn't exist. My prediction is that AI will go the same way. In three to five years, AI photo editing will be as normal as Photoshop. Right now we're in the post-hype hangover, where people react allergically to anything that smells like AI. Is it too early? For businesses, no. For people, yes, for now.
Where the commenters had a point
To be fair, they weren't completely off. A LinkedIn profile photo isn't a product shot of a coffee mug. It's an identity. It's how someone recognizes you at an interview, a coffee meeting, a conference. For that kind of photo, authenticity weighs more. And AI edits have a signature: perfect waves in the hair, an office straight out of a stock library, lighting without a single flaw. The eye catches it in a second, even when the face stays the same.
So my takeaway? For a photo that represents a specific person, less AI and more truth. A plain phone photo by a window is fine. For everything else, it's a different story.
Use AI, skip AI, disclose AI
| Situation | AI edit? | Why |
|---|
| Product photo background, lighting, dust removal | Yes | The product stays true, it just looks better |
| Banner variants, formats, crops | Yes | Saves time, promises nothing |
| Profile photo, team photo, "about us" | Better not | It's an identity; people will recognize you by it |
| Service results (before/after, haircut, treatment, renovation) | Never | You're promising something the customer won't get |
| People, emotions and story in an ad | Carefully, and disclose | This is where people spot AI and punish it most |
| An entire ad made with AI | Disclose, and have a reason | Without a reason it looks like cutting corners on the customer |
For a beauty salon: AI can clean up the background of a salon photo or create ten Instagram banner variants. But a "before and after treatment" photo has to be real, without a single AI pixel. That's your best advertising and your biggest risk. One exposed edit and the trust is gone.
For a corporation: same logic, bigger scale. AI in production (backgrounds, formats, localization): yes. AI in story and emotion: only when you can defend it and disclose it. Otherwise you end up like the Christmas ad that lasted three days.
What about the law?
Since August 2, 2026, Article 50 of the EU AI Act requires businesses to disclose content that was artificially generated or manipulated in a way that could pass as real (so-called deepfakes). Exactly how this applies to light edits of a product or profile photo isn't fully clear yet. I'm not a lawyer, and this isn't legal advice. A sensible rule, though: if someone could believe they're seeing reality and they aren't, disclose it. In the US, the FTC's long-standing rules on deceptive advertising point the same way.
Expert perspective: "People don't punish AI. They punish the feeling of being fooled. Photoshop went through this twenty years ago and nobody cares about it today. AI will go the same way, but the brands that use it badly will lose trust before it becomes normal," says Michal Krčmář, founder of Don Marketer and a CMO with nearly two decades of experience in global marketing.
The takeaway
Before you hit "edit with AI" next time, ask yourself one question: am I promising something that isn't true? If not, go ahead. If yes, skip it or disclose it. And if someone under your post starts debating the photo instead of the point, that's fine. At least you know they're reading.
The public doesn't trust AI. Marketers, on the other hand, trust it far too much. More on that next time.
And now the ending I promised. The post where people debated whether my friend's face was real brought her four real job offers. In a small region. None is signed yet, but one is 99% there. She's thrilled. So am I.
Five percent of people debated the photo. The rest helped. Hate is louder, but help is stronger. That's the point I wanted to make all along.
And if you're wondering how much to put into brand and trust, so one unlucky photo doesn't undo years of work, the Don Marketer calculator gives you the brand : performance split for your business in 2 minutes.
Sources: DesignRush: McDonald's AI Holiday Ad Backlash (December 2025, incl. criticism of Coca-Cola's AI ads); Washington State University: Using the term 'artificial intelligence' in product descriptions reduces purchase intentions (Cicek, Gursoy, Lu, Journal of Hospitality Marketing & Management, 2024); TechCrunch: Meta changes its label from 'Made with AI' to 'AI info' (July 2024); Osborne Clarke: Advertisers must include label identifying retouched photographs (France, Oct 1, 2017); Cooley: EU AI Act transparency obligations take effect 2 August 2026; the author's own experience (LinkedIn post, October 2026, comments quoted anonymously).
