🤖🍴AI: The Flavour of the Month — and the One Designing the Next?

- Tweet0
- 0
- Pin It 0
- 0
- Subscribe to RSS
Subscribe

Subject: 🤖🍴AI: The Flavour of the Month — and the One Designing the Next?

Was this email forwarded to you?

Click here to join our community and make sure you follow us on LinkedIn

With 58% of global (and 64% of U.S.) VC money now flowing into AI, it’s no longer the elephant in the room — it is the room. But while it’s easy to see AI trimming food waste through smarter logistics or menu prediction, a bigger question looms: can AI actually create food people will love to eat?

AI Can Now “Taste” Sweetness — Without Taking a Bite

Scientists are training models to predict flavor before a molecule ever touches a tongue. Platforms like TastePepAI and the aptly (😂) named FART (Flavor Analysis and Recognition Transformer) can now forecast sweetness, bitterness, umami, and sourness purely from chemical structures.

FART, trained on 15,000 compounds, reached ~91% accuracy in taste prediction. TastePepAI has already uncovered dozens of promising peptides — speeding up the hunt for natural flavor enhancers and safe sugar substitutes.

In short: AI can now skip months of lab work and go straight to molecules that taste good and do good.

🤔 Will Consumers Trust AI’s Taste Buds?

While labs are racing ahead, consumer psychology is lagging. Most research focuses on how AI develops flavors, not how humans feel about eating them.

A recent study explored “cognitive trust” (AI’s perceived competence) and “affective trust” (emotional comfort). Both are shaky when food feels… algorithmically generated.

Still, some brands are experimenting boldly.

Beijing, February 2024: MUJI launched AI-developed French fries, created after 3 trillion recipe simulations. The packaging proudly declares:

“This is the flavor AI believes humans will enjoy.”

The fries — in “Southeast Asian,” “Chinese,” and “Western” varieties — got people talking. Unfortunately, the “Chinese” flavor reminded many consumers of herbal medicine or braised meat. Turns out, even AI can fall for cultural clichés.

Elsewhere, Luckin Coffee lets AI design its orange latte poster, Yili Group uses AI for packaging, Juewei Foods deploys AI agents for precision personalized marketing to 120 million members with 3.1x higher sales, and Coca-Cola’s “Y3000” flavor claimed AI involvement from concept to marketing. But most of these applications stop at creative support. When it comes to taste, AI’s output still feels more like a novelty than a breakthrough.

Figure 1: The launch of Juewei Foods' AI Membership Smart Assistant, in partnership with Tencent Smart Retail

In China’s booming ready-to-eat meal market, where freshness and wok aroma reign supreme, AI-generated flavors often struggle to evoke that essential human touch.

🧩 AI’s Data Drought: The Real Barrier to Food Innovation

We dug into the issue, speaking with industry veterans and AI practitioners. The takeaway: AI could revolutionize food manufacturing — if only it weren’t starving for quality data.

Most factory tweaks and recipe adjustments aren’t logged systematically, leaving AI with no feedback loops. For complex processes like twin-screw extrusion (key to plant-based protein texture), this means poor model accuracy.

Recent meta-research exposed just how shaky the field is:

81% of AI flavor models aren’t reproducible,
63% lack real-world validation,
80% rely on inaccessible data.

Without rich, transparent datasets reflecting real processing conditions, AI risks serving up generic guesses instead of actionable insights. It’s like trying to improve your cooking — without ever tasting the dish.

The New Face of AI in Food: Redesigning Work, Not Replacing People

The smartest food companies are realizing that AI’s biggest value isn’t in replacing humans — it’s in restructuring how they work.

At Yili's farms, 6,500 tons of milk flow through daily while smart collars monitor each cow's vital signs around the clock. The result? Quality control that starts at the source, not the factory gate — with protein content stable above 3.4% while reducing bacterial counts to 1/20 of EU standards. The factory floor tells the same story. Yili's 40 smart factories use predictive algorithms to forecast regional demand and auto-adjust production schedules, dramatically improving 24-hour delivery rates. On the consumer end, Yili rolled out AI shopping assistants that coach sales reps in real time, delivering 15.7% more product clicks and 26% higher order rates without losing the human connection that drives trust.

Figure 2: Automated system for milking cows at Yili's smart farm

Global food services giant, Sodexo, applies AI from farm to fork. By merging IoT sensors with AI analytics, it shifted maintenance work from reactive checklists to predictive upkeep — preventing failures before they happen. The result: engineers spend less time on routine tasks and more time solving real problems.

In both cases, AI doesn’t steal jobs; it redesigns them. It amplifies human judgment, sharpens decisions, and creates smarter, more sustainable systems.

So, What’s on the Menu?

AI is already transforming how food is imagined, produced, and sold. It’s decoding flavor molecules, testing billions of recipes, and reshaping workflows — yet it still struggles to recreate the art, emotion, and serendipity of human taste.

That may be its ultimate challenge: not making food that’s efficient, but making food that feels alive.

For now, the flavour of the month is AI — but the real question is: will we still want a second serving?

Save the Date: Future Food Asia 10th Anniversary

We’re continuing this conversation offline. Future Food Asia returns 12–13 May 2026 where investors, startups and innovators meet to rethink how AI is reshaping Asia’s food and agri landscape.

📍 More details soon but trust us, you’ll want this one on your calendar.

Discover all the public sources that informed this newsletter here, plus a layer of ID Capital’s proprietary research — our analysis, data, and field notes that bring it all together.

Subscribe

Subject: 🤖🍴AI: The Flavour of the Month — and the One Designing the Next?