Products You May Like
Robots are finding their way into virtually every industry requiring physical labor, and agriculture is no exception. Tech startups and university research labs have spent the last several years working to create robots capable of tending to fragile salad greens and weeding massive crop fields. While these robots are physically effective, the AI systems that power them are sometimes confused by the vast expanses of green they deal with; weeds become virtually indistinguishable from young or leafy crops.
Researchers at the University of Copenhagen in Denmark have proposed an eye-catching solution: turning crops blue. By making genetic alterations to wheat, maize, and other common crops, scientists can make them stand out to the robots tasked with taking care of them. This, they say, makes agricultural robots less likely to pluck a desirable crop from the soil by mistake.
In a paper for Trends in Plant Science, a trio of biologists suggest that genetically engineering crops to produce natural pigments would improve the efficiency of remote sensing. They focus on anthocyanins (the class of flavonoids that give blueberries, blackberries, red grapes, plums, and other darkly-colored crops their pigment) and carotenoids (which give carrots, tomatoes, pumpkins, and other orange-tinted crops their characteristic color). While farmers have bred crops to produce anthocyanins and carotenoids reliably before, these researchers suggest using modern gene editing techniques to circumvent generations of trial and error.
What an AI-powered weeding implement sees, per the startup FarmWise.
Credit: FarmWise
These techniques could also make it easier for the agricultural industry to pivot around climate change-related challenges. The biologists argue that many of the high-yielding crops we rely on today are susceptible to stresses exacerbated by climate change: drought, atmospheric warming, sharp changes among local ecosystems, and so on. Meanwhile, many wild (AKA non-domesticated) plants are tolerant to these increasingly harsh conditions. Although working to domesticate wild plants—or breed them with domesticated crops—could help farms spread their metaphorical eggs among more baskets, those, too, might be difficult for agricultural robots to distinguish from weeds. The researchers suggest introducing anthocyanins and carotenoids during the domestication and hybridization process, allowing those intentionally grown plants to be easily viewed among undesirables.
In a paywalled NewScientist article, Charles Fox, an AI engineer who was not involved in the above paper, says the biologists’ proposal is “controversial.” Crops that are known to be genetically engineered still get a bad rap, which could make the “huge amount of effort” associated with color crop-engineering far less worthwhile. Still, the biologists argue that they would primarily focus on the crops’ colors, leaving their taste and nutritional integrity alone.