Artificial intelligence is a topic that’s getting a lot of recent attention, thanks to language learning models like OpenAI’s ChatGPT. While a chatbot that uses others’ writing to build responses is novel, we wondered what it might think about how it could be used in irrigation practices. We asked ChatGPT a few questions about AI, automation and ag irrigation, and then brought those responses to industry experts (Justin Gibson, PhD, senior product manager, global agricultural software for Lindsay Corporation, Omaha, Nebraska, and Derek Heeren, PhD, PE, associate professor and irrigation engineer in the department of biological systems engineering for University of Nebraska-Lincoln, Lincoln, Nebraska) to grade them.
ChatGPT: Yes, artificial intelligence can manage irrigation schedules for crops. Advanced AI systems can utilize a combination of sensors and data analytics to monitor various parameters like soil moisture levels, weather forecasts and other environmental factors. Based on this data, the AI can optimize when and how much to irrigate. However, it’s crucial to integrate this AI technology with existing infrastructure and maintain some level of human oversight to ensure optimal results.
Heeren: My interpretation is that it’s pretty close, but it’s a little too optimistic. ChatGPT does an amazing job of searching all the stuff that’s on the web and synthesizing it, so a lot of the concepts here are correct. But I think a lot of the literature on the topic says “This is what we’re trying to do,” but it takes it the next step and says “Oh, AI can manage irrigation.” But the puzzle pieces are on the right track.
Gibson: This is a good answer. If we’re talking about AI in ag, we’ve been using a subset of AI called machine learning for about a decade. In my mind, there’s a number of really great machine learning techniques or applications that optimize or tweak parameters in an algorithm. This is closer to what we call “supervised learning.” It’s great advice to keep a human in the loop here too. While an algorithm can arrive at the right recommendation based on the data it has, if one or more of those data streams are wrong (for example, a broken weather station), its advice might not match reality.
ChatGPT: Growers can significantly improve their irrigation efficiency by integrating automation into their systems. Automated irrigation systems use sensors to continuously monitor soil moisture levels and other environmental conditions. By pairing this real-time data with predictive analytics, the system can automatically adjust irrigation schedules. This eliminates guesswork, reduces water waste and helps in providing the right amount of water to the crops at the right time.
Gibson: This isn’t a bad response, but I see the benefits of automation being realized less on the predictive analytics and more on the decision support. For example, shutting down pivots automatically after a good 1.5-inch rain — it might take half a day before folks can walk the field, check all of their rain gauges and then shut the panel off. If a grower has access to cellular telemetry-connected weather stations and center pivots, we can shut down automatically and save the water that would’ve been applied while folks were out checking their field.
ChatGPT: Automated systems frequently track a variety of data including soil moisture levels, environmental conditions such as temperature and humidity, water flow rates and sometimes even indicators of crop health. Growers can use this information to make informed decisions about when and how much to irrigate. The data also allows for continuous monitoring of the irrigation strategy’s effectiveness and can be used to generate reports that inform long-term planning.
Heeren: It misses monitoring the system itself, that’s a key thing for understanding “Where’s my system in the field? How’s it performing?” and accounting for that in the thought process. One of the comical sentences is where it says that “growers can use this information to make informed decisions” about irrigation. We’re talking about automation — I thought the computer was making decisions!
Gibson: This feels like an answer from a graduate student who got put on the spot and didn’t really know what to say! It just kind of threw out everything as a possible answer.
ChatGPT: A good first step for growers interested in incorporating automation into their irrigation system is to conduct a comprehensive assessment of their current setup. This involves identifying inefficiencies, understanding the specific needs of their crops and determining the scalability of existing systems. After this, growers can begin by implementing basic automated components like timers and soil moisture sensors, and then gradually incorporate more advanced features.
Gibson: A good first step for automation is connecting your devices to the cloud. If you want to do automation, you’re going to want to keep an eye on your systems as they’re running. If we can know when you’re irrigating and how much we can start building a model and bringing automation onto the farm. But I can’t tell you how much to irrigate tomorrow if I don’t know how much you did yesterday.
Heeren: If I’m thinking about the reality of time constraints for a producer; they’re probably not going to do a comprehensive assessment to determine every inefficiency, right? That’s where the role of industry comes in. So the first step might be for the grower to pick up the phone and call their irrigation dealer and say “Does this make sense?”
ChatGPT: For additional guidance, growers should consider connecting with irrigation consultants who can offer expert advice tailored to their specific needs. Extension services from agricultural colleges can also provide valuable insights. Moreover, networking with other farmers who have successfully implemented automation can offer practical tips and recommendations. Industry organizations often offer resources and platforms to connect with experts in the field as well.
Heeren: This is a university perspective! ChatGPT doesn’t show up in the local coffee shop where the farmers talk about what works and what doesn’t work, but it gets the university literature. We’ve got really good extension educators in Nebraska, but in general, their bread and butter is providing the scientific concepts, not getting into a particular product. So the dealer is who you talk to about “Does this product work or not?”
Gibson: This isn’t a bad response, it just feels like a response from someone who doesn’t work in the industry. If you want to learn more about opportunities to automate or bring in data, having a dealer there who can say, “Maybe you’ve got bad cellular connections, so we’re going to have to set up some short distance radio or satellite solutions,” or asking about a pivot — having someone who understands the physical components of the pivot is going to be able to help you get the solutions to capture the data you need to later use for machine learning.
ChatGPT: I hope these answers provide you with the comprehensive information you’re seeking.
The information provided is based on general principles of agricultural irrigation and automation, compiled from the data and patterns I’ve been trained on. My training data includes a wide range of sources, such as academic papers, websites and other texts, up to January 2022. However, it’s important to note that while I aim to provide accurate and useful information, my responses should be used as a starting point and not as a substitute for professional advice, especially for specialized topics like irrigation management.
Heeren: I think this is a really helpful starting point. Sometimes if I have a problem I’m trying to figure out, I’m not sure who to ask or what my real question is yet. ChatGPT’s a little bit like talking to my friends at the coffee shop to help me understand what the problem is. It gives me some vocabulary, some basic ideas. From there, I can talk to someone who’s knowledgeable and figure it out.
