When you open your refrigerator after a long day and feel nothing at all, there’s a subtle unease. Not hunger, not inspiration — just a vague sense of defeat. A half-used block of cheese, some tomatoes going soft near the back, a jar of something you bought for a recipe you never actually made. Most people close the door and order something instead. The food stays. Waste builds up. Additionally, it is rarely discussed.
This pattern, which food researchers sometimes refer to as “fridge blindness,” is more widespread than it may seem. Unilever’s mayonnaise brand, Hellmann’s, developed a whole product around the issue. Meal Reveal, which was created in collaboration with Google Cloud and introduced in early 2024, allows users to point their phone’s camera at the contents of their refrigerator and get recipe recommendations based on what’s actually inside. It’s a fairly straightforward concept. It’s difficult to describe, but it’s strangely satisfying to watch it operate—like the refrigerator finally responding.
This is supported by Google Cloud’s Vertex AI platform, which can identify ingredients from a few uploaded photos or a brief video scan. It’s not flawless. Sometimes it misidentifies what’s hidden behind a milk carton or confuses a zucchini for a cucumber. However, it’s accurate enough to be truly helpful, setting it apart from the majority of wellness apps that solely rely on manual food logging, a practice that, as numerous studies have shown, very few people sustain over time.
The move away from recipe matching and toward something more nutritionally significant seems more recent. Platforms like NutriSnap are already attempting this — analysing photographic data of food to deliver real-time breakdowns of calories, macros, and micronutrients. The ambition is a system that doesn’t just tell you what to cook with your leftover carrots, but tells you whether your fridge, as a whole, is actually supporting your health. A nutritional audit of sorts. Whether that’s useful or slightly intrusive probably depends on the person.
It’s still unclear whether the average household is ready to invite that level of scrutiny into their kitchen. A version of this technology that offers instant feedback, reduced waste, and improved meals feels empowering. Another version, which has a somewhat panopticon-like feel to it, logs, scores, and silently evaluates each food option. Companies framing this as a food waste solution seem to be landing better with the public than those leaning hard on the health tracking angle. It matters how it is framed.
With the announcement of its Food AI platform for smart refrigerators at CES 2020, Samsung took the lead in this field. LG had an internet-connected refrigerator even earlier, in 1999 — though that mostly meant it had a screen on the door. Software has always advanced more quickly than actual adoption, and hardware has always advanced more quickly than software. What’s different now is that the camera is already in everyone’s pocket, and the AI is genuinely good enough to make a scan feel effortless rather than fiddly.
The case for the environment is subtly strong. According to WRAP data, the UK alone discards about 2.9 million potatoes and 1.4 million tomatoes every day. Those aren’t abstract statistics — that’s the bag of potatoes you bought on Tuesday sitting in a drawer going green by Sunday. The overall impact is significant if an AI scan encourages even a small percentage of households to use their possessions. It could be very important.
How seamless the experience remains will determine whether this turns into a kitchen habit or a forgotten app. The ideal version of this technology should feel less like conducting a health audit and more like having a sensible person standing next to you, simply pointing out what you already have.
