Should Agriculture Rely on AI?

Artificial Intelligence Takes on Food Waste Challenges

Could artificial intelligence help us stop wasting food? Scientists from Florida State and Oregon State Universities believe AI is the next step in reducing the amount of food tossed in the trash every year due to over ripeness. Using an imaging application, the collaborative team of scientists hopes to equip both retailers and consumers with a tool to help them know when to sell or consume food items based on ripeness.

Despite a sliver of skepticism, I’ve dabbled in the world of artificial intelligence myself the last few months. I’ve asked a couple of questions along the way. One, if I relinquish work to AI, can the program do the task better than I can? Two, does AI give me a benefit, for example, time, information or a new course of action?

I’m considering the same questions as I look at artificial intelligence in agriculture, particularly when it comes to food waste. 

Food Waste Statistics

You can find plenty of statistics on food waste. For all of our fretting over the world’s ability to produce enough food we manage to throw $400 billion in editable items in the trash every year in the United States. Current numbers suggest that worldwide one-third of food produced for human consumption is wasted annually. According to the National Center for Biotechnology Information 61% of food waste occurs inside the home, 26% in restaurants and 13% in food retail establishments.

The core question is, is there a way to ensure that more of the food we’ve already produced makes it into human bellies rather than landfills, and is relying on artificial intelligence the best way for achieving this goal?  If we have a real strategy to respond to this question it's a win for the environment, food industry profit margins, consumer budgets and food security.

Is AI Better than the Human Eye?

Going back to my criteria for using AI I have to wonder, can artificial intelligence pull better fruits and veggies out of the produce section than I can using my own scanning abilities? According to research conducted by the university professors building the smart-phone app, AI might have a better chance of selecting a perfectly ripe avocado than I do.

In fact, the application uses 1,400 images of avocados to assess exactly how firm the avocado I hold in my hand is which should give me an idea of how ripe it is. In the firmness test the smartphone app is 92% accurate. Asking it to assess internal quality is a little trickier, but the artificial intelligence app still manages to be 84% accurate. I’ve never charted my accuracy on avocado selection but I might just start.

At 92% accuracy I could be persuaded to let AI do the work on this one. And if its accuracy extended into other produce options like melons, leafy greens and potatoes I can see how my grocery shopping confidence would increase. Avocados are a pretty common item to assess poorly at the super market, but the university professors believe they are only the beginning for AI shopping assistance.

If we take it a step further, you and I might not even have to worry about choosing between ripe or rotten avocados at all if the retailers themselves utilize the same AI technology. If only the freshest, most ripe avocados are on the stand to begin with then our odds of purchasing the perfect avocado could increase beyond 92%.  Could artificial intelligence allow a supermarket to have a ripe bin and an under ripe bin? If it could, it would certainly give consumers the information they need to purchase what they need today or what they need for next week.

Will Artificial Intelligence Help Agriculture Save Time and Money?

My second criteria for using AI is, does it afford me some kind of benefit I would miss out on if I didn’t use it, is an appropriate question to ask here as well. My guacamole will taste better if all of the avocados are perfectly ripe, but is that enough of a benefit?

The average American throws about $800 worth of food in the trash every year. To put it in perspective, that is around 215 meals.  Artificial Intelligence can’t put that money back into my pocket, but overtime it could make me a more efficient shopper and meal planner.  If we are talking about making informed decisions then AI smartphone apps offering me real time quality assessment mean I’m a little more in tune with the food I consume and can make adjustments to my weekly purchases based on what is in season, what is at peak ripeness and maybe even where it is worth spending extra money.

Retailers could be the ones to experience a greater benefit. We’ve seen promising AI technology along supply chains when it comes to identifying inefficiencies, categorizing consumer trends and resource management.  It's addressing labor challenges and providing real-time analytics. In short, artificial intelligence is becoming foundational to decision making in agriculture and food systems. 

Can the same smartphone app that helps me pick the best avocado also help a grocery store manage their inventory to the point where they are buying less and selling more? Scanning individual avocados, melons, or lettuce bunches on a smartphone so the retailer knows which items to put out now and which can be held back until next week sounds both tedious and highly beneficial to profit margins. It is not a stretch to think that the smartphone technology underway at university research labs could be laying the foundation for broader AI technology able to assess large quantities of produce quickly as they are offloaded into supermarkets and again as they are shelved or stored. The food retail industry loses about $50 billion per year on food waste primarily from the fresh produce category. Just like the household  grocery budget, AI can’t put $50 billion back in retailer’s pockets directly, but it can offer the kind of data that trickles down into more efficient inventory decisions which do have a direct effect on profit margins.