The short answer
Filipinos are calling AI-generated menus an “auto-pass” for two stacked reasons. First, AI fatigue: we see so much AI content online that we spot the style instantly, and seeing it on a restaurant tarp feels like AI following us offline. Second, trust. Jollibee and McDonald’s earned decades of trust, so their enhanced photos don’t fool anyone anymore. With a small business we’ve never tried, the photo is all we have, and a too-perfect AI photo makes the risk impossible to judge. So we don’t take it.
Lately I’ve seen a lot of Filipinos hating on AI-generated menus and posters. They’re even saying “auto-pass,” as in they won’t even consider eating there anymore. They’re sick of the AI design, and sick of the unrealistic, too-good-to-be-true food photos.
But here’s my question: food from Jollibee and McDonald’s has never exactly looked like it does on the menu either. So why are we so angry when small businesses do it? It turns out there’s a real psychological reason our brains work this way.
I made a short video about this too, if you’d rather watch the quick version first.
Why does every AI menu look the same?
Part of the anger is simple fatigue. We see so much AI content online that we can now spot it instantly: the photo art style, the fonts, that painted design under the text. So when it shows up on a restaurant’s tarp, the reaction is “oh no, not again.” It feels like AI is following us offline.
And there’s a reason all these menus look identical. Everyone is using ChatGPT or Gemini the same way. They upload photos of their food, then type: “Create a menu for my restaurant. Make the food look delicious. Our brand colors are white and red.” If it looks good enough, they print it or post it on Facebook. You know who else did exactly that? Every single one of their competitors.
There’s a technical name for why that happens. Generative models are trained to predict the most statistically average, popular visual patterns, so similar prompts converge on the same default fonts, line weights, and vector icons. The business consequence is worse than looking generic: in a crowded food market, visual sameness commoditizes you. If your menu looks identical to five other Facebook sellers, customers assume you share the same supplier, or worse, that you’re a copy-paste scam page, and you end up competing purely on price.
The too-perfect photos trigger something deeper, a visual uncanny valley: when something looks almost real, the hyper-symmetrical textures, unnatural gloss, and plastic lighting produce a subtle unease. And there’s an over-beautification penalty on top of it. When food looks impossibly glossy, buyers instinctively infer the seller is compensating, that the food isn’t good enough to advertise for itself. It’s the same instinct as seeing a dating profile photo that looks too beautiful and thinking, no way this is real.
Why do we forgive Jollibee but not the food cart?
Jollibee and McDonald’s have spent decades earning our trust. We’ve ordered the Chickenjoy and the fries so many times that no matter how much bigger the chicken looks on the menu, or how much fuller that medium fries container looks, we’re no longer disappointed. We already expect it won’t look like that. When we order from Jollibee, we’re not really buying what’s in the photo. We’re buying a taste we already know. The photo is just a reminder.
With a small business we’ve never tried, the photo isn’t a reminder. It’s the only thing we have to go on. So when that photo looks too perfect, too obviously AI, our brain starts doing the math. With Jollibee, we know how far the real thing is from the picture. With a small seller, we have no idea. It could be a little off, or nothing like it at all. And when we can’t tell how big the risk is, we don’t take it. That’s the auto-pass.
You want proof? Think about when people rant about Jollibee online. It’s usually when they launch a new product. In that moment, Jollibee is in the same position as a small seller: we base our expectations on the photo, and when the real thing looks nothing like the ad, we feel cheated. This “expectation versus reality” problem isn’t new to advertising at all. AI just made everyone more aware and more critical of what they see.
What happens to sales when a small business uses AI photos?
One Filipino food SME shared what happened when they trialed AI-assisted photo enhancement, just brightness and saturation adjustments, nothing heavy. Sales were great the first week. Then customers started saying the polished photos made the dishes look overly idealized, a mismatch with the actual product. By week two, even with only light AI refinement, sales dropped. Customers could tell, partly because social media platforms now label AI-touched photos automatically. Some customers even requested non-AI photos, doubling the upload workload. After two weeks, the business reverted to raw, unedited photos. Their conclusion: customers read real photos as transparency, and transparency as trust.
Another micro business found the line that works: they let AI lay out the poster, but never let it alter the food photo itself, and they post original photos alongside for reference. Customers receive exactly what they see, so the trust holds even though the design is AI-made.
That’s really where the line sits: deception. Professionally shot food photos have always used tricks, and nobody riots over good lighting. But when the quantity or the presentation in the photo is a lie, people feel cheated whether AI was involved or not. Enhance plating, clean up the background, make the food look appetizing. Anything that crosses into budol is just lying to the customer.
Should small businesses use AI at all?
At the end of the day, AI is only as good as the person using it. If you know what good design looks like, you can make menus that look great without deceiving anyone or scaring them away. And that’s a learnable skill, if you put in the effort. There’s also still real value in learning to take your own photos; you’d be amazed what natural light can do for your menu items.
If you don’t put in that effort, and you just type “make this look good” like 99% of other owners did, then you’ll have to be okay with AI making you look like everybody else.