Walk into a Jusco supermarket in Yamato, a small city near Tokyo, Japan, and you can glimpse the future of meat. In addition to a conventional bar code, each steak package sports its own ID number. Type the number into the computer sitting on a nearby table, and up pops information about the cow the steak came from: a scanned copy of its negative test result for mad-cow disease and, in case you are interested, its breed and sex, its date of slaughter, and the name of the producer. At some Japanese meat-counter displays, you’ll even see a picture of the family that raised the animal.
All this information is available because the steaks come from Japanese cattle that have been individually tracked from birth, generally with radio frequency identification (RFID) tags; each of the cows has an ID number correlated with a database entry that documents its birth date, medical history, and movements from feedlot to slaughter, and the results of mandatory mad-cow tests. At slaughter, the ID numbers, and all data linked to those numbers, are passed on to individual boxes of meat.
It’s the future of the meat industry, perhaps, but still a far cry from how things are done in the United States, where few of the 96 million cattle scattered across 800,000 ranches and feedlots are subject to comprehensive electronic digital record keeping. Indeed, fewer than 5 percent are electronically tracked from birth to slaughter, and even then, the identifying information is generally lost during the processing of the meat. “Certainly, at the steak level, you literally don’t have any idea of where that animal might have come from,” says Geoffrey Dahl, a professor of animal science at the University of Illinois at Urbana-Champaign.
All that could be about to change, however, as the U.S. meat industry faces immense pressures to update its antiquated practices. The wake-up call was the finding of mad-cow disease in Washington State late last year. The federal government took four days to determine where the sick Holstein came from, which didn’t exactly inspire confidence that future outbreaks-whether of mad cow or any other disease-could be readily contained. The scare decimated the $3.3 billion annual U.S. beef export market, as more than a dozen countries-the biggest in terms of consumption being Japan-slammed their doors shut to U.S. beef.
In response, the U.S. Department of Agriculture is pressing beef producers to voluntarily adopt new technologies so that by next year, any individual animal can be traced back to its birthplace within 48 hours. A pending bill in the U.S. Congress would mandate a still undefined system to electronically track cows from birth to slaughter. Meanwhile, Japanese and other foreign buyers are demanding proof that all cattle have been tested for mad-cow disease before they reopen their markets to U.S. beef.
The U.S. meat industry is consequently facing an extreme infotech makeover involving everything from RFID tags to retinal-scanning identification to global-positioning tracking technology-and even DNA testing. It’s a makeover that’s long overdue. “Today, a lot of this is done by gosh and by golly,” says Gary Acromite, chief information officer at Greeley, CO-based Swift, the nation’s third-largest meatpacker. “The traceability process is intended to take an old, old, old, old process that is very manual and modernize it with emerging technologies.”
The coming rush to trace meat could provide benefits that would extend far beyond ensuring that beef is safe from mad-cow disease. Indeed, for many sectors of the $95 billion livestock industry-not just beef but pork, lamb, and other meats-better tracking could provide a way to document to consumers that products are from animals raised on, say, organic feed, or that they have been tested for a wide variety of diseases. Moreover, once the meat industry gains the ability to track animals from birth to market, it will be able to determine which produced the best cuts and use that information to optimize breeding, veterinary care, and feeding practices. “From my point of view, it’s a no-brainer,” says Ray Goldberg, a professor of agribusiness at Harvard University’s John F. Kennedy School of Government.
Eyes Wide Open
Bringing the digital revolution to the farm is a daunting chore. For starters, most U.S. cattle producers are still mom-and-pop operations selling fewer than 500 calves every year. Margins are slim, and farmers tend to be skeptics about new technology. “Despite what Congress is saying, this is not something where we wake up one morning and say, Wow, why don’t we roll out traceability to 800,000 beef producers across the country?’” says Mark Armentrout, chief operating officer of AgInfoLink, a Longmont, CO, company that sells radio frequency identification tags and software.
The seemingly Herculean task is, however, somewhat eased by the fact that a few large corporations dominate the meat-processing business. Three companies-Tyson, Cargill, and Swift-together process more than two-thirds of U.S. beef. And the lessons from the mad-cow outbreak in the United Kingdom during the 1980s and ’90s are never far from the minds of officials at these firms. The British outbreak infected 200,000 cows and resulted in the precautionary killing of some 4.5 million more. Following reports of deaths from the human version of the disease, called variant Creutzfeldt-Jakob disease, markets throughout the world closed to British exports. Ultimately, the U.K. outbreak resulted in the deaths of some 150 people, and some export markets remain closed to U.K. farmers to this day. U.S. meat producers are also well aware that what finally quenched the epidemic and reassured consumers was not only increased testing and a ban on dangerous feeding practices but a rigorous tracking system. These days, every cow raised in the United Kingdom is tracked for life with a bar-coded “passport.”
The United States is still far from having such a system in place. But large meat-processing companies like Swift are getting started. Swift’s novel approach is based on retinal scanning to record the unique vascular patterns in cows’ eyes. A handheld device is placed in front of the animal’s eye, and a few seconds later, the pattern is recorded and the animal positively identified. The cows don’t mind: “The first thing they do is they open their eyes real wide. They kind of look at it and stop,” says Bruce Golden, CEO of a small Colorado company called Optibrand, which developed the system and supplied it to Swift.
The retinal scanning is used on a portion of Swift’s own feedlots to keep track of the animals it raises-roughly a third of the 5,400 cattle it slaughters every day at its Greeley packinghouse. In a number of ways, the technology offers advantages over RFID tags, which occasionally get lost, can be switched by unscrupulous producers, and pose at least some risk of winding up stuck in a steak. Retinal scanning allows Swift to document where its cows were raised, and though it is just one step in tracking each animal’s medical history, it is already giving the company a marketing edge. “Even traceability back to lots of 20 or 30 animals has helped us in the marketplace,” says Swift’s Acromite.
But identifying an animal is only the first part of a tracking system. The real benefit comes from building a database that contains information such as birth date, identity of the animal’s “sire” and “dam” (as parents are known), its weight at various stages of life, dates and descriptions of medical treatments or hormone injections, and sales prices at auctions. “The real key here is that we are betting that in the not-too-distant future, the consumer, the government, and the world will require individual-animal traceability,” Acromite says. And while there’s no clear technology mandate or standard just yet, he says, the company is also integrating RFID tag readers and bar code scanners into its system in order to be ready “no matter what the feedlot guy, the government, or the industry throws at us.”
As a final touch, the handheld scanner developed by Swift and Optibrand includes a GPS receiver. Eventually, if the technology becomes widely used, an animal’s location will be logged in the database each time its eye is scanned. GPS coordinates could be correlated with locations such as the ranch where it was born, the feedlot where it grew up, any auction houses it might have passed through, and the slaughterhouse. This location information would be critical to rapidly identifying the herd mates of a sick animal. Consider what happened in the case of mad-cow disease in Washington State. An inspector noted that an animal appeared sick when it arrived for slaughter and pulled it out for testing. When the test came back positive, it took four days to learn where the sick animal came from and to identify its herd mates-not all of whom were accounted for. But if a sick cow shows up at Swift’s gate, and it’s been tracked from birth, such epidemiology could be accomplished almost instantaneously.
Though retinal scanning may be an important piece of the meat-tracking puzzle, it has an obvious disadvantage: the eye needs to be connected to the rest of the animal. Once the animal is cut apart-“disassembled,” in industry parlance-the retinal scan is no longer useful. Neither, for that matter, is an RFID ear tag. The one surefire identifier is DNA, which can be used to trace any animal part anywhere in the production process-and let consumers know where, exactly, their dinner came from.
This advanced, and still pricey, technology is finding its initial applications in North America’s $13 billion pork industry. Though pork producers might not have to worry about mad-cow disease-there is no equivalent mad-pig disease-they face their own safety worries, as well as growing consumer demand for products from pigs raised in healthy and disease-free environments. For producers hoping to market their products as premium pork, it is critical to document just where the meat came from.
Maple Leaf Foods-Canada’s largest producer of pork-plans by this November to market pork products that are tracked by the industry’s first DNA-based meat-tracking technology. It includes a DNA test and a database that will allow pork products to be conclusively traced from a store shelf back to their origins. Initially, the products will be sold in Japan, where consumers have proven willing to pay extra for reliable information on meat origins. At a current cost of 40 Canadian dollars per test, the technology is clearly not suitable for routine testing of packages of meats, but it could be used to spot-check shipments of pork to assure consumers and stores that they originated on Canadian farms and are safe. “We want to reposition the Canadian pig industry in terms of food safety. We can prove that this piece of meat in Tokyo came from Canada, and from Maple Leaf, and from a certain food production system,” says John Webb, the Toronto-based company’s director of genetics and science.
The technology, developed by Pyxis Genomics of Chicago, consists of a simple test that determines the presence of a set of genetic markers known to be common to a mother and all her offspring. Once a mother’s DNA fingerprint is in a database, a tissue sample from a piece of meat can be tested to see if it matches. It’s a technology particularly well suited to the arithmetic of pig breeding. A sow can produce 50 to 70 piglets in a lifetime, so a single blood sample from a mother pig provides a way to conclusively verify the source of an enormous quantity of chops, loins, and bacon.
A key element of the Maple Leaf system is a database under development by IBM Canada. At first, the database will contain limited information; if a sample is tested and a match found, it will simply show that the pork came from the offspring of a particular sow at a certain Maple Leaf supplier’s farm. But the database is being constructed to handle more detailed information about pigs, such as their birth dates, weight gains, medical treatments, and breeding history. “It’s extremely scalable, so all the pigs in the world could be in this database, and it still could run quickly,” points out Susan Wilkinson, who heads the Maple Leaf database project as an associate partner with IBM’s business consulting service in Toronto.
Maple Leaf says it hopes to eventually use a single DNA test that can be used to not only trace meat but also screen it for pathogens like E. coli and salmonella. “That’s the dream ticket, and a lot of people are working on it,” Webb says. He’s optimistic that such a test will be available in five years, helping to enhance the safety of pork supplies. “DNA is very much the platform for the future,” he contends.
But while DNA testing is an exciting prospect, it remains a niche opportunity for the marketing of high-end food products. For now, it’s far too expensive to play a role in most of the U.S. and Canadian pork industry, never mind at North America’s 800,000 cattle farms. And even for more mundane and readily available technologies like RFID tags, the gap between optimistic promises and the realities on the farm and feedlots means improving the safety of the nation’s meat supplies will be a stiff challenge.
Consider the day-to-day realities faced by Ed Greiman. Greiman owns a modest feedlot in the northern-Iowa town of Garner that each year fattens 2,400 cattle for slaughter. In his barn, an eight-year-old Fujitsu laptop shares a wooden shelf with assorted hand tools. Duct tape binds the computer’s cracked case, and plastic welding compound keeps the display from falling off. But the machine flickers to life, and AgInfoLink’s cattle management program, called BeefLink, lights up the screen. The program helps manage the growth and health of individual animals, identified by means of RFID tags on their ears. With the program’s aid, Greiman can do things like provide custom medical treatment and cull poor weight-gainers early.
One late-winter day, Greiman opens a metal gate and wades into a cattle pen to demonstrate the benefits of the technology. In the pen, 80 skittish cattle stomp on cornhusks as they go about their life’s work: munching on a mixture of hay, cow corn, and a high-calorie, high-protein yellow slurry. Greiman corrals a brown-and-white cow through a steel cattle chute then yanks a lever that pins the animal’s neck between two steel bars. Then he waves an RFID tag reader past a quarter-sized white tag on the cow’s ear.
This one is number 1565, and the database tells a story that might otherwise have gone unheard. On November 30, her temperature was a slightly feverish 39.4 C, and she was given some antibiotics. On December 2, it was still 39.4 C, and she got more antibiotics. Still later in the month, the fever persisted. Most troublesome to the bottom line, though, was that the ailing number 1565 was simply not gaining weight. It happens all the time: some animals are better suited for feedlot life than others. Any additional investment-antibiotics, hormones, yellow slurry-won’t produce a higher carcass weight. The ear tag helped Greiman make a clear decision: animal number 1565 would be on the next truck to the packinghouse.
But Ed Greiman is also frustrated by what happens on either end of his process. Fewer than 10 percent of the 990,000 Iowa-raised beef calves supplied to feedlots like Greiman’s each year carry any individual records. That makes it nearly impossible for Greiman to custom-order calves with consistent or predictable genetics. Instead, he must work with whatever walks through his gate. On the outgoing side, the data Greiman dutifully collects dies along with the animals inside a Tama, IA, slaughterhouse. Without a large-scale, integrated network for gathering genetic and disease information on meat throughout the food chain, well-intentioned efforts like Greiman’s are little more than duct tape on a food-safety disaster waiting to happen.
It’s a long way from the realities of Garner, IA, to the high-tech Jusco supermarket in Japan. But today’s meat industry is increasingly global, with shoppers often buying products raised and processed halfway around the world. Those global consumers, in turn, are increasingly concerned about mad-cow disease, bacterial contamination, and other safety and quality issues. In short, consumers want to know where their meat is from, and they want guarantees that it is safe. And the industry faces looming regulatory and market pressures to adopt information technology.
As the information age dawns on U.S. meat producers-whether modest operations like Greiman’s or huge corporations like Swift-the differences between how things are done in Iowa and Tokyo could eventually dissolve. So while high-tech tracking like Jusco’s might seem exotic to U.S. consumers, it may be coming to a supermarket near you. And if it does, a family buying a steak might see a photo of a smiling Ed Greiman, as well as a certificate documenting that its dinner is safe.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.