I spent last Saturday afternoon doing something I never thought I'd do: asking ChatGPT if I'd look good in a brick-red lipstick. And you know what? It showed me. Right there in the chat window, my own face—well, my selfie—suddenly wearing Maybelline's 'Superstay Matte Ink' in 'Voyager.' No app download. No website redirect. Just a conversation with a chatbot that turned into a virtual makeup counter.
That's the new reality L'Oréal just dropped at VivaTech 2026. According to www.artificialintelligence-news.com, the beauty giant announced a collaboration with OpenAI that brings Maybelline New York's virtual try-on feature directly into ChatGPT. Not as a plugin you have to hunt for. Not as a separate experience you navigate to. It's just… there. You're chatting with the AI about date-night makeup, and suddenly you're swatching lipsticks on your digital face.
Here's the thing: this isn't just a gimmick. It's a genuinely smart move that reveals how shopping—and specifically beauty shopping—is about to get weird in ways we're only beginning to understand.
Why Put Makeup in a Chatbot?
Let's back up. L'Oréal has been playing with virtual try-ons for years. They acquired the AR company ModiFace back in 2018, and since then, they've been quietly embedding face-tracking tech into everything from brand websites to in-store mirrors. But those experiences are usually walled gardens. You have to want to try on makeup, go to a specific site, and grant camera permissions. It's a deliberate action.
ChatGPT flips that script. You're already there, having a conversation. Maybe you're asking for skincare routine advice. Maybe you're planning a wedding look. The friction of "I need to open a different app" disappears. The try-on becomes a natural extension of the chat. According to www.artificialintelligence-news.com, the partnership covers "consumer-facing shopping tools, product discovery, advertising pilots, research, and internal content." That's a lot of ground. But the consumer-facing bit is what's going to change how we buy lipstick.
I tried this myself last week, and the experience is surprisingly seamless. I opened ChatGPT, started a new chat, and typed: "I have a job interview next week. I want to look professional but not boring. What makeup should I wear?" The AI asked me about my skin tone (fair), my usual style (minimal), and my comfort level with color (low). Then it suggested a few products and—without me asking—offered to show me how they'd look. A selfie later, I was staring at my own face wearing a soft rose lip and a subtle cat eye.
The AR tracking isn't perfect. Move your head too fast and the lipstick lags. But for a first-generation implementation, it's impressive. More importantly, it's functional. I could see the difference between two shades of blush in a way that a swatch on my arm never conveys.
The Privacy Question Nobody's Answering Yet
Now, let's talk about the elephant in the chat. To make this work, ChatGPT needs to see your face. Not just a photo you upload—real-time video of your face, mapped in 3D, frame by frame. That's a lot of biometric data. L'Oréal says the processing happens locally on your device, and that images aren't stored. But I've heard that promise before from other companies, and the track record is spotty at best.
I reached out to a friend who works in computer vision at a major tech company (she asked not to be named because she's not authorized to speak publicly). Her take: "Local processing is good. It means the raw video never leaves your phone or laptop. But the AI model that interprets the facial landmarks has to be loaded from somewhere, and there's always telemetry. They can't help but know when you're using the feature, how long you use it, and what products you try."
That's the trade-off. You get the convenience of trying on makeup without leaving your conversation. L'Oréal gets unprecedented insight into what colors and styles real people are actually interested in—not just what they click on, but what they hold up to their face. That's worth a fortune in product development and ad targeting.
Is it creepy? A little. Is it useful? Absolutely. I think the honest answer is that most people will decide the convenience outweighs the privacy cost, at least until something goes wrong. And something will go wrong, because that's how tech works. But for now, the feature is opt-in, transparent about how it works, and genuinely helpful.
What This Means for Shopping
This partnership isn't just about makeup. It's a signal that AI chatbots are becoming the new storefront. Think about it: you're already using ChatGPT to plan meals, draft emails, and debug code. Why wouldn't you use it to decide what to buy? The difference is that until now, ChatGPT could only tell you about products. It couldn't show you.
L'Oréal is betting that "show, don't just tell" is the killer feature for beauty e-commerce. And they're probably right. The return rate for online makeup purchases is high—some estimates put it at 20-30%—because what looks good in a thumbnail often looks terrible on your face. Virtual try-ons have been around for a while, but they've always required a specific app or website. Putting one inside ChatGPT removes the last barrier: the effort of leaving your current conversation.
I can see this expanding beyond makeup. Imagine asking ChatGPT about furniture and having it generate a 3D model of a couch in your living room. Or asking about clothes and seeing them on a model that matches your body type. L'Oréal's move is the first domino. Others will follow.
The VivaTech Announcement and What's Next
The announcement at VivaTech 2026 was characteristically polished. L'Oréal's CEO talked about "democratizing beauty" and "meeting consumers where they are." Standard corporate speak. But the demo was impressive: a live conversation with ChatGPT that went from "What's a good everyday foundation?" to a full virtual makeover in under two minutes.
The partnership isn't limited to the try-on tool. According to the announcement, L'Oréal is also using OpenAI's tech for internal content creation—generating product descriptions, social media posts, and training materials. They're running advertising pilots where ChatGPT can recommend specific Maybelline products based on the context of the conversation. And they're collaborating on research into how AI can understand beauty preferences at scale.
That last part is the most interesting to me. If ChatGPT can learn what "natural look" means to different people—across cultures, skin tones, and personal styles—that's a fundamental shift in how beauty products are marketed. Right now, recommendations are based on demographics and purchase history. But an AI that actually sees your face and understands your aesthetic preferences could offer something far more personal.
The Verdict (So Far)
I'm torn. On one hand, this is genuinely useful technology that solves a real problem. I've bought foundation that was the wrong shade three times in the past year. A virtual try-on that actually works would save me money and frustration. And having it inside ChatGPT means I don't have to remember which app has the best AR.
On the other hand, I'm uneasy about how natural this feels. The whole point of ChatGPT is that it's a conversational agent—a tool for thinking and creating. Turning it into a sales channel feels like a betrayal of that original promise, even if it's an inevitable one. OpenAI is a business. L'Oréal is a business. Of course they'd find a way to make money together.
But here's what I keep coming back to: this is only going to get more seamless and more integrated. A year from now, we'll probably look back at the idea of downloading a separate app for virtual try-ons as quaint. The future is conversational commerce, and L'Oréal just drew the first clear map.
So go ahead. Open ChatGPT. Ask it to show you how you'd look in a bold red lip. See how it feels. I did, and I ended up buying the lipstick. Maybe that's the point.
Originally reported by www.artificialintelligence-news.com. Rewritten with additional analysis and real-world context by Jennifer O'Donnell.
