Your Sushi May Be Getting Smarter

'Small data' technology could finally replace food products' expiration dates.

Issei Kato/Reuters

Every year, some 48 million people in the United States get sick from something they ate. And thousands of them die from these foodborne illnesses, according to the Centers for Disease Control and Prevention.

But the food that can make you sickest often doesn't even look or smell tainted. Simply giving a food an expiration—or use-by—date doesn't do much to protect people from bacteria like salmonella and e. Coli. After all, it's not just time that spoils perishables; it's also temperature.

Americans end up throwing away tons of food—thousands of dollars' worth per grocery store each day, according to one report—in an abundance of caution, and then many of them get sick anyway.

One of the key factors contributing to this ongoing public-health problem is the question of which way the data is flowing. The systems we have in place now to track food safety are largely centralized: Huge agencies like the CDC and FDA collect information, track sickness as it's reported, and disseminate crucial public-safety notices.

But what if individual food items had smart labels that gave consumers the information—beyond simple expiration dates—to determine whether something is safe to eat from the moment they pick it up at the store? Thinfilm, for example, makes paper-thin electronic labels that are bendable and rewritable. Its CEO, Davor Sutija, says there's value in offering more item-by-item information without relying on centralized infrastructure to make sense of it.

Increasingly we hear talk about smart refrigerators linked to temperature-sensitive smart labels that might jointly tell you when to toss a particular item. (The Atlantic recently wrote about such technologies here.) And it's easy to imagine how useful it would be to connect smart labels with a larger database—the FDA's trove of food recalls, for instance. But Sutija says consumers increasingly want smart technologies that don't plug into that kind of networked infrastructure.

"We're not offering big data, we're offering small data," Sutija told me. What he means is data that relies not on a larger database, but rather on specific information that a consumer can glean just by looking at a given product.

Imagine, for instance, a tray of salmon sashimi with an electronic label that has tracked the temperature of the fish at each stage in the supply chain—from the warehouse in Norway to the local processing plant to the plane to the truck to the grocery store and, eventually, to the display case where customers decide whether to buy it. While a best-by date might hint at when the dish was prepared, an electronic label gives a much more precise backstory of the fish's path from the boat to your chopsticks.

Sure, some people might love the idea of integrating those labels with a centralized database of recalls and setting up related push notifications so they receive up-to-the-minute alerts about food they purchased that might not be safe. But others, especially as people grow weary about sharing personal data, might prefer smart data that's first "sensed in the periphery" and then moved to the center, rather than the other way around.

This is, as Sutija describes it, "data that doesn't have to be aggregated with all of the data in the cloud." And that is why, he claims, it's part of an overall system that is good for consumers. A single label costs "in the dimes," he points out, and not only is there "a cost advantage for those offering it," there is also "a privacy advantage for those using it."

"This is not Big Brother," Sutija says. "This is allowing smart objects to be the agent of the consumer, to sense things about the environment around them."