From AI to robot butchers, automating the meat packing industry is not cut-and-dried

The Canadian agriculture industry has introduced many new technologies over the last decade, from data systems advising farmers when to spray, seed and harvest their fields to roboticmilking machines and AI to create the ideal facilities to grow crickets

Meat processing plants are advancing, too — just not nearly as fast.

Artificial intelligence, automation and robotics are reshaping many parts of the economy, and experts say the meat-processing sector could benefit from improved technology to overcome labour shortages, improve food safety and remain competitive against alternatives like plant-based proteins.

Still, change has been slow.

“We are laggards, in many ways,” said Andreas Liris, the chief information and technology officer at Maple Leaf Foods. “We’re making small investments, but definitely behind the curve compared to some other sectors.”

 

 

 “With the rise of lab-grown meat and plant-based proteins, the stakes are high for the meat-processing sector to become more efficient”

 

The manufacturing sector, including automotive plants, has relied on robotics for several years, but putting a bolt in a car is very different than working with meat, since no two animals are the same.

One of the challenges is replicating the human eye and touch. So far, robotic butchers aren’t able to make precise cuts and can also struggle to accurately tell the difference between skin, fat, bone and meat in chicken and turkey facilities.

“Super difficult, but getting better, though. Robots are becoming more tactile,” said Liris.

Tough enough

Companies are making strides, such as using robotics to transport heavy animal carcasses within a facility, to stack and move boxes for delivery and to optimize transportation loads to reduce the amount of trucking. 

In recent years, Maple Leaf has begun using sensors on trucks to track the temperature and humidity for flocks of chickens from a farm to a facility. The technology helps improve animal welfare and meat quality, said Liris.

Robotic equipment still isn’t able to properly decipher between skin, meat, bones and fat in poultry. (Bryan Eneas/CBC)

Several Canadian startups are trying to develop solutions to modernize the sector, including Waterloo-based P&P Optica, which uses hyperspectral imaging and artificial intelligence to understand food chemistry as the meat moves along the production line.

After developing sophisticated imaging technology, the next challenge was ensuring the equipment was durable enough to function in cool temperatures and withstand being blasted with hot, soapy, high-pressure water during the sanitizing of a facility.

“How do you put a very fragile optical system into a manufacturing plant that you have to clean everyday?” said Olga Pawluczyk, chief executive of P&P Optica. “You have to clean it in such a way that you don’t have salmonella, listeria, E. coli growing in the plant. It’s not a gentle environment, let me put it that way.”

The company now has its equipment installed in 10 meat packing facilities in North America. The imaging technology detects and measures quantities of protein, water and fat, while also noticing any contaminants, such as plastic debris that could break off from a conveyer belt.

Labour woes

The meat processing sector has always been labour-intensive and filling those positions is a constant challenge. The low-paying jobs are often undesirable, as workers on the factory floor repeat the same movements with knives and saws to debone and slice up livestock.

Introducing AI, automation and robotics in meat-processing facilities is difficult, experts say, largely because the raw material is variable, since no two animals are the same. (Jean-François Blanchet/Radio-Canada)

During the early stages of the COVID-19 pandemic, meat-processing facilities were hit hard, as the virus easily spread between workers standing nearly shoulder to shoulder. The pandemic highlighted the industry’s reliance on those workers, as numerous plant shutdowns interrupted the meat supply to grocers and drove up prices.

“One of the pressure points that the industry is feeling, which is labour, could be alleviated by investments in technology,” said Mike von Massow, a food economist at the University of Guelph.   

“Those problems are going to become more acute over time,” he said, describing how reducing the labour intensity of the sector could help moderate food price increases, although many other factors also impact price tags on grocery store shelves.

Some in the industry see technology as the answer to alleviate labour pressures, while others envision robots assisting workers, not replacing them.

One technology that could reduce the labour requirements of a facility is being developed by Michael Ngadi, a bioresource engineering professor at McGill University and the founder of MatrixSpec Solutions. For the last 10 years, he’s developed a tool to detect how much fat or marbling is in a piece of pork. North Americans generally prefer lower fat, while some Asian countries favour higher fat.

Currently, workers are trained to assess the amount of fat on the processing line, said Ngadi, but a combination of machine learning and AI could be more reliable.

“If you have someone who has worked eight hours or so just looking at pork products, at some point, naturally, the efficiency of that person might actually decrease. Or just simply someone having a headache or a bad day and you can have those issues,” said Ngadi.

Better cutting, better profits

MatrixSpec Solutions is one of several startups receiving funding from the Canadian Agri-Food Automation and Intelligence Network (CAAIN), a not-for-profit supported by the federal government.

Another project receiving financial support is Mode40, which aims to improve profit margins for meat packers, while also reducing their energy use.

One of the technologies of the Steinbach, Man.-based company is using data and AI to determine how to optimally cool an animal carcass in the most energy-efficient way.

In addition, the wrong temperature can increase the difficulty of making precise cuts, said Mode40 CEO Cameron Bergen, explaining how meat can be stiff if it’s too cold and slippery if it’s too warm. 

Making the wrong cut can result in less meat selling at a premium (like steak or prime rib) and more of it sold at much less value (in ground beef or sausage).

Robotic equipment has to be able to make precise cuts to not leave any meat on a bone, while also being strong enough to move large pieces of meat. (Brian Higgins/CBC)

“It may only add up to quarters for that meat, but when you’re in a facility that processes one million or two million or more animals a year, those quarters per meat add up to dollars per carcass. And those dollars per carcass add up to significant impacts for the industry,” he said.

Technological advances throughout the meat processing sector are important for the much broader agriculture industry and for the national economy, said von Massow. Canadian beef exports alone were worth $4.7 billion in 2022.

“With the rise of lab-grown meat and plant-based proteins, the stakes are high for the meat-processing sector to become more efficient,” he said. “The sector is embracing innovation, but so far, he said, the pace of change hasn’t been quick enough.”

“I’m not aware of anywhere where there are companies that have really sort of cutting-edge technology — pardon the pun — in a meat-packing plant,” he said.