AI & RoboticsNews

The meatpacking industry is an incubator for AI, automation, and COVID-19

This article is part of a VB special issue. Read the full series: Automation and jobs in the new normal.


In early spring 2020, Smithfield, Tyson, and other industrial food suppliers warned that upwards of millions of pounds of meat could disappear from the U.S. supply chain as a result of the coronavirus. Although it now appears these fears were overblown or possibly a ploy to bolster exports (excepting pork products like pepperoni), tens of thousands of slaughterhouse workers around the world have tested positive for COVID-19, and more than 90 of them have died from the virus.

As the health crisis stretches on, the threat to meatpacking, meat processing, and distribution center employees has researchers hunting for a new production model. Even with physical distancing protocols and personal protective equipment like face shields and masks, plant closures are looming — and the idea of automation is rapidly gaining ground.

The danger zone

The U.S. meatpacking industry employed nearly 600,000 workers — a large portion of whom are immigrants — at wages averaging $15.92 an hour in 2019. The field has high turnover, and a January 2005 report released by the Government Accountability Office showed that some worksites experience over 100% annual churn. In April 2017, a U.S. Immigration and Customs Enforcement raid exposed slaughterhouses that had knowingly hired — and in some cases trafficked — undocumented employees with the promise of steady income.

Meat processing is dangerous work. The rate of cumulative trauma injuries — serious physical injuries from repeated or prolonged activities — is the highest of any U.S. industry, at about 33 times the national average. According to federal statistics, nearly one out of 10 meatpacking workers suffers a cumulative trauma injury every year, up from one in four workers just over 20 years ago.

The unrelenting pursuit of speed and profit is likely to blame. The more animals are slaughtered per hour, the less it costs to process each one. In 1976, the typical line speed in a U.S. slaughterhouse was roughly 175 cattle per hour. By 2001, that had climbed to around 400. (One Tyson-owned plant in Holcomb, Kansas reportedly slaughters up to 6,000 head of cattle per day.) The U.S. Department of Agriculture (DOA) imposes line speed restrictions on a per-industry basis, but — citing an internal rule change in 2018 — it granted a record number of waivers to poultry and swine processors earlier this year. Plants owned by Tyson and Wayne Farms, among others, were allowed to operate lines at 175 birds per minute instead of 140. In response, activist groups filed a lawsuit against the DOA, arguing the waivers made conditions more dangerous for workers.

On the line, hundreds of people stand in close proximity, wielding sharp materials as carcasses hung on hooks from overhead chains move toward them. Boston Consulting Group estimates meat processors employ 3.2 workers per 1,000 square feet of manufacturing space, or 3 times the national average for manufacturers.

Lacerations are common — workers stab themselves or someone nearby — as are accidents involving power tools, conveyor belts, falling carcasses, and slippery floors. Repetitive motion injuries can lead to lifelong impairments as workers repeat motions throughout their shift, making the same knife cut 10,000 times a day or lifting the same weight every few seconds. Occupational Safety and Health Administration (OSHA) 2014 data showed that repetitive motion injuries among beef and pork processing workers were 7 times that in other industries.

Robots to the rescue?

There’s a limit to what humans — even pushed to the limit — can physically do. This is certainly true in the meatpacking industry, which began to introduce automated machinery and robots as early as the 1960s.

While robots can’t address every task on the meat processing line, they’re increasingly able to perform the bulk of them — from packaging and sealing to cutting thoraxes and extracting viscera. Machines can scan, weigh, and measure carcasses to eviscerate them “intelligently,” with the more sophisticated models planning blade trajectories for cutting, separating meat from carcasses and boning them out.

Dunedin, New Zealand-based Scott Automation is one of the largest meat processing robot providers in the world, with operations across five continents and customers in over 80 countries. In collaboration with partners like Meat & Livestock Australia, it develops and supplies machines like the Primal System for lamb, which uses computer vision to create a 3D map of bones within lamb forequarters, middles, and hindquarters for precision cutting that factors in height and angle measurements.

For customers operating at scale, there’s Scott Automation’s automated boning system, which comprises six machines that transfer meat from one to the next in sequence. It separates up to 12 carcasses a minute into three sections using the Primal System for guidance, processing the forequarter, middle, and hindquarter before removing the knuckle tip from the hind leg. A reconfigurable middle system locates the spinal cord holes at either end of the saddle section, using a combination of vacuum and compressed air to remove spinal materials while a chine station bones the rack saddle, adding up to five grams to the yield per carcass.

Scott Automation sales director Chris Hopkins told VentureBeat via email that while some of the company’s machines reduce the need for manual labor, none fully replaces human workers. He pointed out that available space in already tight facilities is often a challenge and that the systems’ primary value derives from the accuracy of their cuts, which increases carcass profits.

“We … work closely with the processor companies to help equip their staff with the skills required to support and maintain the technology. We do not believe you would ever (or want to) fully replace all workers,” Hopkins said. “There is some thinking that automation and robotics can reduce the reliance on or the total number of people required to operate a processing facility, but any solution that does that … is still some years away.”

Tokyo-based Mayekawa agrees with that assessment of meat-processing robots’ capabilities. The meatpacking automation company, which has sold thousands of chicken and turkey leg deboners, neck slitters, neck skin removers, and gizzard openers to companies in over 26 countries, says its products are designed to work alongside workers, rather than stand in for them. Customers who replace manual pork and ham deboning work with its machines realize a 50- 60% workforce reduction in deboning-related tasks, Mayekawa claims.

“We think that automation is vital to maintaining the food supply chain,” Mayekawa engineering manager Shinji Shimamura told VentureBeat. “The pandemic has brought to light the need to further automate wherever possible.”

Mayekawa’s marquee product — Legdas — separates chicken bone and leg meat at a rate of up to 3,000 legs per hour with results “as good as when done by hand.” The company also offers a poultry cutting machine that segments cuts with a touch pen, eliminating the need for workers to carry carcasses to a workbench or use a knife, and a feather bone extractor that automatically measures carcasses and peels bone from meat. (Together, they process 150 to 200 head per hour.) And Mayekawa recently developed Hamdas, a swine deboning machine that uses X-ray technology and AI to identify left and right legs, pick out femur bones and shank bones, and make slittings in up to 500 carcasses per hour with blade-equipped arms.

In some cases, automation in the meatpacking industry has the potential to shift rather than reduce the demand for labor. A crab-processing robot developed by the Canadian Centre for Fisheries Innovation (CCFI) that cuts crabs in half and removes their legs is designed as part of a robotic system to extract meat from the crabs’ shells. It’s a process that’s typically done overseas, but the designers assert it can solve some of the workforce problems in rural Newfoundland fish plants that are attributed to changing demographics.

“Younger people are not being attracted to the industry … If you talk to operators of fish plants today, everybody needs more people,” Bob Verge, managing director of the CCFI, told the Canadian Broadcasting Corporation. “A large part of the labor force in our processing sector now comes from the baby boomer generation. We can’t replace those baby boomers with an equal number of younger people.”

The coronavirus problem

At the urging of regulators, a number of processing plants have implemented infection prevention measures, including surgical masks, fever screening, and barriers encouraging six feet of distance between workers. However, companies complain these measures might slow down production. Indeed, U.S. pork processing dipped 6% year-over-year for the week ending May 30.

Both Scott Automation and Mayekawa claim that some of their machines reduce workers’ risk of coming into contact with the coronavirus. The Centers for Disease Control and Prevention’s latest report notes that workers who are struggling to keep up breathe harder and might have difficulty keeping masks properly positioned on their faces. Other experts theorize the cold temperatures — and aggressive ventilation systems — required to prevent spoilage could be allowing the coronavirus to stay viable for longer.

Scott Automation says most of its robotic systems operate in “exclusion zones” to keep people away from the equipment and maintain safety. There might be as many as two operators, but they’re physically distanced, and machines can be installed with “clean-in-place” technology that automates sterilization. Mayekawa says its Hamdas system optionally sterilizes blades as it processes carcasses. However, the company concedes most of its products require daily maintenance and upkeep, putting the onus on workers to maintain a proper distance and scrub down surfaces.

There’s evidence — albeit anecdotal — that automation technologies have prevented (or at least forestalled) some COVID-19 infections among factory workers. Costco’s high-tech chicken processing plant in Nebraska, which employs two shifts of roughly 400 employees, reported only a single COVID-19 case in mid-April. In Denmark, across all 18 of the Danish Crown’s almost entirely automated meatpacking facilities, fewer than 10 workers have tested positive out of 8,000. And in Michigan, Clemens Food Group’s robotic pork-cutting packaging plant stayed open through May, slowing production only to install new protective equipment.

By comparison, a Smithfield-owned facility in Sioux Falls, South Dakota disclosed hundreds of cases in early April, culminating in the facility’s closing. Tyson was also forced to shutter its Columbus Junction plant after dozens of workers contracted the virus, as were meatpacking operators in Canada, Spain, Ireland, Brazil, and Australia.

Business as usual

Meat processors are likely to further embrace automation for all the reasons mentioned — safety, yield, and reduced labor. Last August, Tyson, which has invested more than $215 million in robotics over the past six years, opened a facility near its headquarters in Springdale, Arkansas to develop automation solutions for its production plants. (In July, the Wall Street Journal reported that Tyson engineers and scientists are developing an automated deboning system to help butcher the nearly 40 million chickens processed each week.) In fall 2015, Brazil-based JBS — the world’s largest meatpacker — acquired a controlling share in Scott Automation. And Pilgrim’s Pride invested over $30 million in automation last year, targeting projects it says are helping its plants run efficiently in the midst of the pandemic.

“We believe in automation, we believe in robotics, and we’re going to continue to move down that path,” Pilgrim’s Pride CEO Jayson Penn told analysts during an April earnings call. “This is something that [even] pre-COVID we’ve been addressing and doing with our facilities, using more automation and more robotics.”

But change is unlikely to occur overnight. Tyson, JBS, Cargill, and other meat giants say robots can’t yet match humans’ ability to disassemble animal carcasses that subtly differ in size and shape. Finer cutting such as trimming fat largely remains in the hands of human workers. A skilled loin boner can efficiently carve a cut of meat like filet mignon without leaving too many scraps that get turned into lower-value products, such as the finely textured beef used in hamburger meat.

And automating even a portion of the line is expensive. While Scott Automation offers sub-$200,000 solutions, it says its customers generally spend in the millions of dollars and don’t anticipate a return on investment for at least a year. “[The pandemic] will raise interest in automation, but I’m unsure if it will accelerate adoption,” Hopkins said. “That will come down to how much the meat processors — our customers — want the equipment and are prepared to invest their resources to achieve a faster adoption.”

This article is part of a VB special issue. Read the full series: Automation and jobs in the new normal.


Author: Kyle Wiggers.
Source: Venturebeat

Related posts
AI & RoboticsNews

H2O.ai improves AI agent accuracy with predictive models

AI & RoboticsNews

Microsoft’s AI agents: 4 insights that could reshape the enterprise landscape

AI & RoboticsNews

Nvidia accelerates Google quantum AI design with quantum physics simulation

DefenseNews

Marine Corps F-35C notches first overseas combat strike

Sign up for our Newsletter and
stay informed!