CAGs respond to May topic: Artificial Intelligence

As agriculture undergoes a transformative evolution, propelled by technological advancements, Artificial Intelligence. (AI) emerges as a pivotal force reshaping the landscape of farming practices worldwide. From precision farming to crop monitoring, AI applications are revolutionizing the way we cultivate, manage, and harvest crops, promising increased efficiency, sustainability, and productivity. In this era of digital agriculture, the integration of AI algorithms and machine learning techniques holds the potential to address pressing challenges such as resource optimization, pest management, and yield prediction, ushering in a new era of smart farming. In exploring the intersection of AI and agriculture, we delve into a realm where innovation meets necessity, offering solutions that not only enhance agricultural processes but also pave the way towards a more resilient and food-secure future.

The paragraph above was generated by ChatGPT, an AI chat bot after it was asked it to “Write an introduction paragraph for answers to questions dealing with Artificial Intelligence in agriculture.” It is almost unbelievable how far this technology has come — and makes you wonder what the future holds. It’s important for agriculture to have a seat at the table, and I’m glad you all were able to gather around to discuss this emerging technology.

Here’s how you humans responded to the May discussion topic on the issue:

What incentives are available and what more are needed to allow farmers to acquire and scale these technologies?

• Many large corporations are already collecting lots of data. (i.e. Fieldview) They give some amount of data analysis in return but aren’t always up front on how they will use our data. If the technologies are made available, they could allow farmers to reduce input costs and chemical use. Cost is a large impediment for significant implementation. (The Cornstalks; ProFILE)
• In a world that is heavily reliant on cutting-edge technology, it seems to stop us in our tracks when a network crashes or the power goes out. How do we comfortably rely on AI to take the person out of the task? There are potential benefits in being able to get more done with fewer people, or in getting tasks done more efficiently with current workforces. If we were to substitute out a person for technology, we’d like to see manual operations able to step in, in case of equipment/technology failures. A key downside we feel would be the potential for these technologies to increase the gap between. (Bringin’ the Bacon!; ProFILE)
• We feel money needs to be more available for this, both for the farmers to be able to afford it and for continued research. (The Moo Crew; ProFILE) 
• The incentives available are open-source material on the internet. “Ag open” Brian Tischler programming with C++, and Arduino boards to outfit GPS on equipment. In the future, having equipment leasable through some program similar to how Livingston County Soil Conservation rented out a no-till drill at a reasonable price — when no-till planting was new. It may require an operator/instructor to be part of the package as the new equipment may not be easy to run and maintain. (Kirk’s Farm Bureau Group; Livingston County)
• Incentives: greater adoption of sustainability practice through AI utilization. Use of AI for EID tags and recording for reports/report findings. Repeatability and reliability of tools will be major factor to adoption. Using AI to help farmers communicate or find and apply for grants and funding has potential benefit. (The CDC; ProFILE)
• We are not aware of any. (AgVentures; Saginaw County)
• We do not know about any incentives available to move us into using AI. We do know that 25 years ago we thought we were high-tech using foam markers for our spray equipment. Little did we know that by today how frequently we would be using GPS on our farms. If AI will save us time and/or money, we may not need other incentives. (Modern Producers; Monroe County)

What can be done to ensure that quality data is fed into AI models to make sure unbiased and accurate predictions are made?

• Make sure users know the data quality is critical. This will be difficult to verify. (The Cornstalks; ProFILE)
• The first step is to identify what the task at hand is trying to accomplish by working with farmers to figure out how this technology could be beneficial, and where this fits into the different pieces of the industry. (Bringin’ the Bacon!; ProFILE)
• Just as the adage goes for farm work: quality in = quality out. The concern is that AI takes from all sources, which may conflict or provide misinformation. Additionally, AI may self-filter and be misleading. (The CDC; ProFILE)
• We recognize that humans will always be biased. We can account for this with making sure we use a large database as much as possible. We also recognize that data predictions aren’t always perfect, and we have to take that error factor / chance into account. (The Moo Crew; ProFILE) 
• Quality data needs to come from primary sources that can be trusted and valid. It may be possible to set up an AI program to check sources. (Kirk’s Farm Bureau Group; Livingston County)
• Yield model — you calibrate it. Trust plus confidentiality issues. (AgVentures; Saginaw County)
• This may be a role of our land-grant institutions. We are afraid government will step in and take on this role. We need someone who has a good “ear” on agriculture to oversee that quality data is being used in AI models. We point to the recent “40B package” as an indication of how little government truly understands agriculture. (Modern Producers; Monroe County) 

What protections need to be in place to protect both data privacy and security with AI?

• Make sure user data is anonymous when entered into a large database. Make sure you know what the program will require of you and how it will use your data. (The Cornstalks; ProFILE)
• What type of infrastructure does AI run on? Among privacy concerns, can this system be protected so that the “answers” AI works off cannot be maliciously altered? (Bringin’ the Bacon!; ProFILE)
• Farm data should always be proprietary and remain in the farmer’s control. Unfortunately, there are greater concerns for hacking and data security. Additionally, transparency of how and where the input and output from the AI machine is sourced and traced — that will help build trust. (The CDC; ProFILE)
• We recognize that data is hackable and that we need to create better ways to scramble the data to prevent as much hacking as possible. We need to do better than what’s currently available. (The Moo Crew; ProFILE) 
• Protection of data privacy and security should be current and updated as soon as possible. (Kirk’s Farm Bureau Group; Livingston County)
• Who needs the information and who will be using it? (AgVentures; Saginaw County)
• We do not know how this protection will be implemented, but are 100% in favor of protections being developed. (Modern Producers; Monroe County)