Artificial Intelligence in the Food Safety Industry

23 May 2019
Louis Corso

Artificial Intelligence in the food safety industry

When I think of artificial intelligence I conjure up images of robots that are highly intelligent and able to make decisions without human intervention. In my mind it is Jarvis from The Avengers movie series. In reality we are far, far away from that type of Artificial Intelligence. Still the future is very exciting and if you come away with one thing from this article I hope it is I want to know more.

What is artificial intelligence?

In computer scienceartificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and animals.


So where are we?

When we say Artificial Intelligence we really mean machine learning. In this article we will look at how a machine or machine learning has the ability to make clear concise accurate decisions based on enough quality data to reduce or eliminate the risk of arriving at incorrect conclusions.  In this case, how can we plant, pick, pack, process and distribute food from farm to table and improve the quality at each point along the way. Improvement in the food supply chain can help increase production, reduce waste and by small incremental improvement, make a significant impact by the time a product reaches the store shelf.

Ok, so when we say AI we are referring to machine learning. i.e. the ability for a machine to learn over time through the acquisition of data. The critical element of this process is over time. Understand that two things have to happen. One, machine learning has to be introduced into the equation with some knowledge based on data already mined and two machines have to be on the job long enough to acquire new data to then make decisions to improve the process. The addition of machine learning alone will not immediately solve problems and there is a substantial cost to introduce these new tools. For the moment let’s put the cost aside, because it is always about the ROI and there could be a substantial ROI.

Business is always looking for new tools and methods to understand how consumers tastes evolve over time.  If you are in the business and chances are you are, then you know that Consumers tastes are extremely diverse and fickle. For that reason and a number of others businesses are turning to new means to not only find ways to please consumers, but to also improve the quality of foods that are being produced.

Imagine that you could improve by just a small amount at each point along the path everything from the raw material through transportation and manufacturing up until the product reaches the consumers hands. Unless you own the means of production of the raw material, distribution, processing and final shipment to the store the cost of machine learning will not be borne by a single business, but by all. Does that mean prices will rise at first and then fall because there is a cost to absorb? Good question! The initial introduction will be at significant cost and not all will be able to afford it. Over time, like all technologies the price will come down and more and more businesses will be able to use the technology.

Cost will definitely be a factor, but will machine learning change the landscape? It should be imperceptible to the consumer because growers, distributors and producers will still need to compete. Over time those who adopt it first will have a significant edge over their competition and as always speed to market will be a differentiator.  And we know that by introducing this new technology into the chain we can improve at every point along the way.

How does it work and why do we need it?
Harvesting data to ensure the food you eat is a huge application for Machine Learning. More than likely in ways that we have not even thought of yet.  Farmers will use machine learning to deploy drones that can map topography and look at crops from above to make determinations about quality in real-time and make adjustments. E.g. how much to water the crops and where, to yield the greatest harvest. Also, machines that will harvest the crops can be made out of alloys so they are not susceptible to bacteria and temperature fluctuations and can be cleaned and sterilized at the end of the day.  More data will come from temperature sensors attached to intelligent shipping to providing optimal environments in transit. Facial recognition software will be deployed to determine if an item is blemished and other sensors can determine if a product is emitting an odor indicating spoilage.

Over time the acquisition of data will result in increasing the quality of the food that will reach your table. Think about it for a moment if we could improve the quality by a small fraction across the entire food chain there would be a major impact by the time the product reaches the store shelf.



One more thing. I hope to write more about Artificial Intelligence or for now Machine Learning algorithms.  If you do decide to keep reading about Machine Learning may I suggestion you grab a copy of Pedro Domingos “The Master Algorithm”.

I got mine from my local library.

Leave a Reply