Senate Agricultural Committee Hearing On Innovations In AI Farming Technology 11-14-23
We'll introduce each of our witnesses now and then turn it to you for your comments. Our first witness is Dr. Mason Earles. He is an assistant professor at the University of California Davis, where he leads a research lab and team focused on AI and agriculture. Dr. Earles is also a co lead for the USDA
funded National AI Institute for Next Generation Food Systems. Prior to joining UC Davis, he worked at Apple as a data science engineer and recently founded an agriculture technology startup called Scout. Our next witness is Dr.
Sanjeev Krishnan Mr. Krishnan is the Chief Investment Officer and founding managing director at S2G Ventures. He has nearly 20 years of experience in sourcing and managing venture and capital equity investments focusing on agriculture and food companies.
Mr. Krishnan is a graduate of the London School of Economics and grew up in Grosse Pointe, Michigan. So we're and the special reason why I'm glad that you're here. He enjoys spending time with his family and friends in their home city of Chicago, Illinois. Thank you so much for being here.
I'm now going to turn to Senator Thune to make the next introduction. Thank you, Madam Chair. And let me just echo what was said earlier by the ranking member, Senator Boozman.
And thank you to you and to him for working in facilitating an extension of the Farm Bill and dealing with the orphan programs. I'm pleased we're able to get that done, although I would say that it is no substitute for a multi year authorization. And I hope we keep our heads down and try and get that accomplished. I do want to welcome José-Marie Griffiths to the panel today. She is President at Dakota State University. I've had the privilege of working with her for several years, and she has testified numerous times in front of the Senate Commerce Committee where she shared her expertise.
And I'm grateful that she is willing to share that expertise today with the Senate Ag Committee. Dr. Griffith has served in presidential appointments to the National Science Board Commission on Libraries and Information Science, and she's recently been appointed to be a member of the National Security Commission on Artificial Intelligence. Because of Dr. Griffiths passion about expanding research and educational opportunities, Dakota State University is a leader in the nation in developing a high quality skilled technology and cyber workforce. And DSU has played a critical part in ensuring that our farmers and ranchers are able to take part in the technological revolution.
Dsu, in partnership with South Dakota State University, launched a joint doctorate program on precision Ag and cyber. This program brings an important focus to the cybersecurity threats the agricultural economy faces. And this is critically important as new technologies like artificial intelligence become more prevalent in the agriculture sector, it's important work, this hearing is important. There is a huge intersection between agriculture, which is our state's number one industry, and technology, all of which can boost yields and incomes for agriculture and make us more competitive in the global marketplace. And so, Dr. Griffiths,
thank you for your great work and welcome here today and I look forward to hearing yours. And thank you to all the panelists for being here to share your expertise with this panel. Thanks. Thank you. Senator Boozman will have the next introduction. Yes. I'm pleased to welcome Dr. Jahmy Hindman,
who's Senior Vice President and Chief Technology Officer at Deere and Company, A position he's held since July 2020. He's responsible for building Deere's tech stack, the company's intuitive to software into an equipment solution made up of hardware and devices embedded software connectivity, data, platforms and applications. It's so complicated, I can't even read. He leads the company's intelligence solutions group, its global network of technology innovation centers and the shared engineer function, Working in both the agriculture and turf and construction and forestry divisions. Dr. Hindman has more than 25 years of advanced technology,
artificial intelligence, product engineering, and manufacturing experience. Most recently, he led the engineering team for Deere's flagship tractor product line. Dr. Hindman holds a Bachelor's degree
in mechanical engineering from Iowa State University, as well as a master's and doctoral degrees in mechanical engineering from the University of Saskatchewan. His doctorate focused on the application of artificial neural networks and heavy equipment applications. He sits on the Industrial Advisory Council for Iowa State University's College of Engineering and leads to the technology leadership and strategy initiative for the US Council and Competitives.
Dr. Hindman, thank you very much for being here. Welcome. And for our last introduction, Senator Ron. Thank you, Madam Chair. Happy to introduce fellow Hoosier, Mr. Todd J. Janzen. He's an attorney and co founder of Janzen Schrader Agricultural Law LLC, a law firm dedicated to serving the needs of farmers, Ag technology providers, Agribusinesses located in Indianapolis, Indiana.
Todd also serves as the administrator for the Ag Data Transparent Project, a national effort to bring transparency to contracts between farmers and technology providers. Todd grew up on a grain farm in Kansas. He served as chairs of the American Bar Association's Agricultural Management Committee in the Indiana State Bar Associations Egg Law Section. In addition to his regular law practice, Todd serves as General Counsel to the Indiana Dairy Producers. Todd publishes a national recognized egg technology blog, the Janzen Egg Tech Blog.
Thank you for being here and the rest of the witnesses. Welcome to all of you. And now we'll turn to Dr. Earl's first for 5 minutes of testimony. And we certainly welcome any other information you wish to give us for the record. Great, thank you. Good morning, Chairwoman Stabenow, ranking member and members of the committee.
I'm happy to be here today to discuss the topic of leveraging technology and artificial intelligence for innovation and American agriculture. My name is Mason Earles, and I'm an assistant professor at the University of California Davis. I'm also a PI and Agriculture Production Cluster lead at the USDA funded National AI Institute for Next Generation Food Systems.
And as mentioned earlier, a co founder of Agtech start up. Prior to joining UC Davis, I worked as a data science engineer at Apple. Before this, I worked on the basics of how plants work, fundamentals of plant physiology. Today, my lab at UC Davis sits at the crossroads of agriculture and artificial intelligence, or AI. And I lead a team of engineers, computer scientists, and biologists who are making AI enabled sensing systems that aim to help agricultural producers manage more precisely, efficiently and sustainably. However, I'm not here just to talk about what we do in my lab, but I want to discuss more broadly the topic of rapidly growing trend in AI and technology, and agriculture and food systems, and the role of research institutions in spurring such innovation.
We're sitting here today because of unprecedented advancements in hardware and software, which have massively expanded the capacity of AI computer programs to learn from complex real world data like what we see in agriculture and food systems. Before going any further, however, I would like to define AI using a relevant example. So put simply, an AI is a computer program that takes in one or more inputs, like an image or an audio recording, or table of data, and outputs some prediction or physical action.
Okay, What do I mean by that? So as an example, let's say we input into our computer program thousands of images taken from a camera mounted on a tractor weeding implement in a carrot field. And we want the AI to predict if an invasive weed such as nightshade is present in the image, which could, in response, trigger a precise herbicide spray application. For instance, we typically train these AI computer programs by showing the many examples of inputs with correct outputs in our carrot versus nightshade example.
This means that a human first categorizes which images do or do not have nightshade, right? Literally clicking yes or no on a screen. The AI computer program then repeatedly tries to predict which images contain nightshade versus carrots, and is penalized for incorrect predictions Once training reaches a desired performance target. This AI can then be used to automate the detection and potentially spraying of nightshade based on input images without the need for human categorization. This is just one of many possible examples of the growing number of current and potential applications of AI, and agriculture and food systems.
At the AI Institute for Next Generation Food Systems, our team of more than 40 researchers across six national institutions, aims to accelerate critical solutions to big challenges in the food supply chain, from crop breeding and farming, food processing and nutrition. Imagine if AI could bring together genomic and sensor data to uncover novel molecular patterns to enable plant readers to discover more flavorful and nutritious strawberries. Now imagine those same strawberries growing in a field where hundreds of cheap paper clip sized soil sensors are measuring nutrient and water stress. And wirelessly sending data to a drone that flies overhead each day. After harvest, these strawberries are transported to a processing facility. And rapidly that rapidly samples wash water to rule out the presence of Ecoli pathogens.
Using AI enable microscopy. Then finally, a consumer could point their phone at a plate which uses AI to estimate the macro and micronutrients of the strawberries and every other ingredient they're about to eat. Critically, the socioeconomic and ethical risks of introducing AI tools across the food supply chain, such as data, privacy and security, and potential effects on labor must be considered as well.
Our researchers and industry partners at the AI Institute for Next Generation Food Systems are investigating each of these topics. Among many others, we see ourselves as one of the world's leaders in research development and commercialization of such, importantly, open sourced AI based food solutions in food and agriculture. We do this through a three prong strategy of multidisciplinary science, industry engagement and workforce development. Such a massive effort and innovation is made possible by more than $20 million in funding provided over five years by USDA's Nifa.
As part of the NSF, NSF's National AI Institutes. In fact, four additional national AA institutes focused on food and agriculture are funded by USDA Nifa, totaling more than 30 research institutions and industry partners across America. And these national AI institutes are working on programs that aim to relieve labor shortages via AI driven robotic harvesters and tree crops. Monitor the health and stress of livestock using AI enabled sensors. And predict climate and crop risk by building AI accelerated models that could eventually be used to precisely control irrigation and nutrient emitters.
Each of these AI institutes is focused on tech transfer and meeting industry needs with partners including dairy producers, soybean farmers, chemical and agricultural machinery producers, among others. Thinking to the future of our workforce in 2021, more than 161,000 undergraduate and graduate computer science degrees were awarded in the US alone. While we don't have exact numbers, a very small fraction of these students end up working in agriculture and food sector. This needs to change. So building on these accomplishments and to keep America as a global leader in agriculture innovation, I strongly encourage the committee to continue and even expand funding for these national AI institutes and among other funding sources provided through the USDA that focus on AI solutions for food and agriculture. I believe that this is how we will accelerate more research innovation and industry collaboration. And create a wider funnel for motivating more computer science and engineering students to solve big challenges in the agricultural sector via AI and new technology. Thank you.
Thank you very much. Next we'll hear from Mr. Kushnan. Thank you, Mr. Krishnan for coming and sharing your thoughts. Chairwoman Stabnau, Ranking member Boozman, and members of the Senate Agriculture Committee.
Thank you for the opportunity to speak to you today. My name is Sanjeev Krishnan. I'm a founding team member and chief investment officer of S2G Ventures.
S2G was founded in 2014 and is now one of the largest investment firms focused on sustainable food and agriculture solutions and technologies. Today, we manage 2 billion of capital and our portfolio includes investments in more than 90 companies. We are based in the Midwest and deploy capital that creates jobs and returns in the heartland. S2G I helped lead a team of more than 40 sector experts and investors focused on identifying and making investments in technologies and entrepreneurs. Nearly all 90 portfolio companies are either headquartered or maintain operations in the United States.
Across our portfolio, we have companies that either operate, manufacture, or distribute in every state, represented by members of the Senate Agriculture Committee. We don't just write checks. We ask what problems need to be solved and seek to understand how individuals experience it. What does the consumer in Iowa see in terms of price and quality? What is a farmer in Michigan or a rancher in Colorado experiencing? How are communities in Mississippi impacted? We explore those leverage points and make targeted investments across the ecosystem. In practice, this might look like identifying a durable consumer trend. We might consider investing in new consumer products to meet this trend.
Companies using machine learning to develop more flavorful and nutrient dense ingredients through improved sea genetics. What farmers might need to plant this hypothetical crop? The infrastructure needed to take crops from farm field to finished goods, less intensive ways to nourish the crop. Tools that enable farmers to measure and profit from those environmental benefits and fintech solutions that derisk the transition to a new crop. My comments today will draw from the system's perspective, focusing on four themes.
One, the journey, the Journey of the American farmers. A remarkable story of embracing innovation and transformation. Technology and artificial intelligence build on this tradition. These technologies and their applications and implications are just beginning to emerge. But they offer a unique tool kit to rapidly accelerate breakthrough solutions and significant per acre value generation.
Today, farmers are drowning in data and not in solutions. We now have aggregated data points from sensors, machinery, and many other sources, improving that data quality and utilizing the data to drive better, faster, and more efficient and precise tech solutions. These solutions will be able to automatically adapt and moderate the impacts of extreme weather, volatile commodity prices, and more. This also represents a new frontier, deriving value from on farm data.
Third, AI can help enhance the foundation of US Agriculture trust and community. This sounds counterintuitive, but at technology, data analytics and AI have an important role to play in strengthening human relationships. For example, AI offers nearly instant power to analyze and identify patterns across massive amounts of historical research and on farm data.
Combined with their real world experience, AI can enable independent ecronomists or certified crop advisors to offer faster, more precise advice and actual insights to farmers. Ai can help make existing a technologies better and more effective and help weed out ineffective approaches. This enables farmers to focus on their resources and time on options that work best for the conditions on their operations. For public policy plays a critical role in developing and scaling critical A tech solutions. Federal loan guarantees and other financing opportunities offer security to the developers of Nacent technology that once at scale offer a public good in the form of improved sustainability or profitability. Financial instruments will continue to be an important area for private, public partnership and innovation.
Improving data quality and sharing will be critical. The public and private sector should work together to avoid duplet work and focus limited resources on filling data gaps. And it'll be critical to protect privacy, support the farmer, and build tools that account for the full diversity of the food and agriculture system. Thank you again to the committee for your leadership in commuting. Today's hearing. I look forward to responding to your questions.
Thank you very much. And now we'll hear from Dr. Griffiths. Welcome. Good morning, Chairwoman Stabanau and ranking Member Boozman. And members of the committee, thank you for this opportunity to testify.
Today, I'm José-Marie Griffiths, President of Dakota State University located in Madison, South Dakota. At our institution, we're focused on training the next generation of professionals in emerging technology fields such as AI, cybersecurity, and quantum computing. Dakota State University is one of only ten universities nationwide to hold all three center of academic excellence in cybersecurity designations from the National Security Agency and Department of Homeland Security.
And we have innovative R&D campus facilities and public private partnership models that empower students to immediately enter the cyber workforce upon graduation. One example is the Dakota State University Applied Research Corporation, which operates and manages the Dakota State University Applied Research Lab. Additionally, we're partnering with South Dakota State University, the state's leading agricultural institution on collaborative research. Through a Precision Cyber Ag Partnership, where South Dakota State University brings the data generated by Precision Technologies and we leverage our cyber and AI expertise. Agriculture has evolved tremendously over the past 100 years.
Innovative technology is now being leveraged to drive farming equipment, predict crop health, optimize yields, and monitor the entire produce supply chain from seed to stomach. The integration of technology and AI stands to shape the future of agriculture with both tremendous benefits and risks amidst the modern cyber threat landscape. Further research is needed to ensure AI can reach its full potential. The committee considers how to deploy AI across agriculture. Dakota State University offers the following recommendations. First, support the increased adoption of AI and its transformative potential in the US agriculture sector.
When embedded in connected systems, AI technologies enable the widespread collection of vast amounts of data from crops and livestock through satellites, drones, sensors, and robots. Analysis of these data can lower costs and improve yields and production. Second, support stronger cybersecurity protections to safeguard the critical infrastructure of the US agriculture industry. Embedding AI into internet connected farming machinery, vehicles, and devices does inherently make systems more vulnerable to cyber attacks. But there are solutions to protect this critical infrastructure powering our national food supply. Third, support the expansion of agricultural research focused on AI to help increase the sustainability of the agricultural industry.
This research can lead to the creation of new technologies and improved policies that will enhance agricultural productivity and resilience. Fourth, intellectual property confidentiality risk is another key consideration. As AI applications are rapidly developed and deployed, IP confidentiality is essential to protect and prioritize the development of leading innovations in the field. Finally, the fifth recommendation is for a heightened concern for cyber and national security that involves the acquisition of land by unfriendly nations, especially in sensitive areas or close proximity to critical infrastructure and agricultural areas, it's crucial to secure our land for the sake of national security. The United States has a critical opportunity to advance the use of AI to further innovate the agriculture sector, while also helping address very real cyber risks and challenges associated with a growing attack surface.
There are crucial steps that academia in partnership with industry and federal agencies can take to ensure the safe, responsible, and effective use of AI. The agriculture industry has been automating and innovating for decades. While the deployment of AI across agriculture is a transformative shift, it's nothing we can't be prepared for. And I'd like to recognize our leaders in South Dakota including Senator Thune, Senator Rounds, Representative Johnson, and Governor Noem for their continued leadership in this area.
Dakota State University looks forward to continued collaboration with the committee to develop policies to advance the safe, responsible, and effective deployment of AI across agriculture. Thank you for your time. Thank you very much, Dr. Hindman. Welcome. Good morning on behalf of Deere's 80,000 employees worldwide. I want to thank the committee for the opportunity to address you here today. John Deere is dedicated to assisting customers in meeting the increasing global need for food, fuel, shelter, and clothing.
We tackle challenges like limited land, water, and rural labor by leveraging technology, including artificial intelligence, to empower growers to achieve higher productivity with fewer resources. This approach allows farmers to accomplish more with less, while improving both their economic and environmental sustainability. Currently connects over 650,000 machines around the world using terrestrial cellular networks. This allows data generated during farming tasks like planting and harvesting, to be sent to the cloud for analysis.
The insights gained from this analysis help optimize the farmers current tasks, such as improving their logistics and contribute to enhancing future farming seasons. A common concern regarding farmers data is ownership and we are unequivocal on this matter. John Deere, customers retain control over their data including how it is collected, stored, processed, and shared. That said, we also believe that the valuable insights that can be derived from this information will play a crucial role in meeting our industry's collective objective of sustaining a growing global population. Farmers use these same connections to deliver data driven instructions back to their machines, such as prescriptions for applying different rates of fertilizer to different parts of a field.
Navigation information used to auto steer machines, and input specifications like seed and fertilizer. These network services are offered to farmers through partnerships with third party companies, allowing farmers to have greater flexibility in choosing the services based on their own preferences and needs. However, it's vitally important to address connectivity challenges in rural areas, including in field connectivity, to fully unlock the benefit of technology for farmers. In addition to connecting our products, we have significantly increased the computing capability embedded within those products. This allows for more advanced control and enables a unique plant level management capability where each plant can be nurtured to achieve its optimal potential. Our self propelled sprayers, for example, feature nine graphical processing units and 36 cameras.
These cameras can scan a distance of 120 feet at a speed of 12 miles per hour. Through artificial intelligence, they analyze images to identified weed pixels, allowing precise herbicide application only where necessary. This C and spray technology is not some futuristic vision, it's already in the field. In 2023, US farmers achieved an impressive 61% reduction in contact herbicide usage across 275,000 acres of corn, soy and cotton. Saving approximately 2 million gallons of herbicide. But reducing herbicide use is just the start of AI's potential in agriculture.
For instance, we have integrated the same graphical processing units with stereo cameras in our autonomous tillage solution. This application of artificial intelligence allows us to identify obstacles in the fields, prompting the fully autonomous tractor and tillage tool to pause and await further instructions from the farmer who may be engaged in other higher value tasks. This solution directly addresses labor scarcity, especially during critical agricultural periods such as harvest and planting in the past two growing seasons.
This AI technology facilitated autonomous operations on approximately 45,000 acres of corn and soy in North America. As rural to urban migration continues, AI powered solutions like this one become even more essential to US farm productivity. Additionally, we leverage the power of AI to train models using our technical assistance data. This enables us to promptly address customer dealer machine issues. Ensuring swift problem resolution, AI allows us to efficiently identify similar issues across the machine population and expedite solutions for affected customers. As a result, the impact duration during crucial agronomic timing windows is significantly reduced.
The future of US agriculture is being built today with tools that enable data driven decision making by farmers. Artificial intelligence plays a crucial role in unlocking the value of that data and turning it into actionable insights in the field. But we need your help. Us farmers would benefit greatly from incentives to help them acquire the precision technology needed to do their jobs more efficiently and sustainably. As you deliberate the upcoming Farm Bill, I urge you to consider such proposals as the Precise Act and the Precision Ag Loan Act that would expand eligibility for USDA conservation and loan programs to include the adoption of precision technology. Further, bills like the Last Acre Act are essential for farmers to fully leverage the benefits of AI and precision technologies. Putting these technologies in the hands of America's farmers not only improves productivity and profitability for growers, But also enables them to produce enough food, fuel, shelter, and clothing to sustain the growing world population.
That benefits us all. Thank you. Thank you very much. And now we ask Mr. Janzen. Welcome. Thank you. Chairwoman Stabanau, Ranking member Boozman. And members of the committee.
My name is Todd Janzen. I'm an attorney at the law firm of Janzen Schrader Agricultural Law LLC based in Indianapolis, Indiana. And we serve the needs of America's Farmers, agribusinesses, and also Ag tech providers. I'm also here today because I grew up on a farm in South central Kansas and so I have agriculture at my roots. I'd like to make three points here today about agricultural technology, the digitalization of farming, and artificial intelligence. First, there are many ways that farmers already interfacing with digital technologies.
Today, we have everything from farm management information systems, which provide a suite of services to farmers in exchange for a collection of data and allow farmers to analyze their decisions and make informed decisions based upon all that data. On the other end of a spectrum, we have remote sensors and items that are very specific about performing one task. But these are also connected to the Internet and often referred to as Internet of Things devices. And they allow farmers to remotely monitor what's going on on the farm. And then of course, there's everything in between from aerial imagery, satellite imagery, connected machines, and even connected livestock on the farm.
All of these tools share one common denominator, which is that they all collect a lot of data from farmers. Farmers are a sensitive group when it comes to talking about sharing their data and they have a good reason to be so. Secondly, I want to talk about what are some of the reasons that farmers are reluctant to share data with companies that want to offer these digital tools. When farmers are pulled, it's almost always the same, three things bubble to the surface and I'd say the first is a lack of trust in a lot of these platforms.
Farmers just don't know what happens after the data moves to these cloud based platforms. Secondly, privacy concerns. Obviously, this is proprietary data for a lot of farmers and so it is their livelihood and so they want to know that it's protected. Third, from my standpoint, what I also see is a lot of overly complex technical agreements that make it difficult for farmers to understand exactly what it is they're giving up. As far as the data, much of my work has been done to try and alleviate these concerns in the technology space as a number of companies want to move into the agricultural area. I've done this through a project called the Ag Data Transparent Organization.
What Ag Data Transparent does is certify companies that work to show that they are transparent with how they are collecting and using and storing farmers data. To date, this Ag Data Transparent organization has certified over 40 companies and there are still many more that need to be certified and there's much work to do. I'd like to also talk just a couple of minutes about artificial intelligence and how that is arriving on the farm. I like to think of it in different buckets. On one end, we have narrow AI, or narrow artificial intelligence, which takes a number of data points to make a single informed decision about something like, is this a weed or is this a valuable crop? On the other end of the spectrum, we have the general AI, which takes a lot of data from different sources and is focused on trying to mimic human behavior. An example might be analyzing a whole catalog of different seeds that are available and then applying those to a specific farmers fields for a specific area based upon the weather predictions for that year to suggest this is the ideal crop to plant for you.
With all of these though, of course, there's more data collection that remains a concern for protection. Farmers should know how their data is going to be used. They should know when they sign up for AI platforms, if it's going to be used to train these platforms and what that means for them. Finally, I'd like to offer just three policy considerations based upon my work as an attorney in this sector.
First, I think that any policy should focus on leveling the playing field and not stifling innovation because this is such an innovative sector. Second, when I think trust is lacking, then transparency becomes even more important. If anyone is collecting data and they don't have that farmer trust, they have to be extremely transparent with how they're using it. Finally, any platform that uses agricultural data should try to return and equal or greater value of that data back to the farmer in the resulting product. I would say to you that all three of these recommendations are true, whether it is private industry collecting data or whether or not it is the government collecting data for use in some farm program.
Thank you very much. I look forward to the questions today. Well, thank you to all of you for your thoughtful testimony and raising, I think, really important issues. Let me start with Dr. Earles.
The world was a very different place five years ago when we wrote the last Farm Bill. And from your position now, through your leadership, when you see firsthand how quickly everything is changing, how is our understanding of artificial intelligence in the agriculture sector different than five years ago? Where are we going specifically? Could you talk more specifically about what we need to be paying attention to in doing? Sure. Yes, that's a great question. When I think of what's changed in the last five years, I think the biggest change has been for actually software developers. And what I mean by that, this has upstream consequences for everybody that's going to consume the products from AI. But what that means is it's a lot faster to go from nothing to product today than it was five years ago.
And I'm talking about, for example, today, it may only take a day to build an AI model that counts destructive versus beneficial moths on a piece of paper from images, right? Whereas five years ago, a smaller group of people, of skilled individuals, software developers, would have taken them much longer to do this. What that ends up meaning is that we get products, more products to market faster for agricultural AI startups and industry. What hasn't changed though, I think, is that AI models still require data and they require human supervision.
And what I mean by that, as I mentioned earlier, someone has to go in and punch in a computer what they see. That person is an expert, right? An agricultural expert. And those are still in high demand, but low supply. Being able to connect those agricultural experts to the software development process is something that is changing quickly, but still needs to change faster in the next five years. I think you'd asked what I see going into the next five years as well. Even in the last week or two weeks, we've seen a huge change in how AI and humans communicate.
And so I'm going to give a quick example of this and what I did last night. So I sat down and I took my phone out, opened up an app, and I said, you mentioned you generated your comments through AI. I asked it, I said, I'm a soybean farmer outside of Des Moines, Iowa. I have some small worm that's infesting my crop. I'm near a river. What do you think it is? And it came back with five different possible responses within 3 seconds.
And it said, I might be able to help you figure this out if you show me an image first. I did this with my voice. I said this to the AI.
I responded with its voice and text. It then gave me an option to show it an image. This is just like talking to a person, right? I show it an image of what I know is to be this certain type of worm, and it correctly identifies this type of worm to me. This is how we communicate with AI's is changing rapidly and they have soon to be role, if not already, role of being AI advisors. I think this is something we need to think a lot about, is what are the recommendations that they are making and how are we training those models such that they're giving reliable, robust recommendations? And how are we regulating what's coming out of the models? Which may end up being easier than thinking of how farm advisors are actually doing, but we also need to take that into consideration. Thank you very much. And Mr. Krishnan,
when we look at these innovations, of course none of them matter if they're just in the lab. They have to be out in the hands of farmers. And to do that, we'll need a partnership between researchers and the private sector with a willing and able workforce also. So could you speak to what you hear from others in the private sector about their reluctance to invest in agricultural technology and how can we use these advances to bring new high paying jobs to our states? Thank you. So given the need to feed a population of 10 billion globally and America's leadership, there's increase in interest from the venture capital community in this space.
Ten years ago was less than 1 billion. At its peak, we hit 12 billion year to day. We venture capital is invested close to 6 billion in the sector. I will say though, a lot of venture capital investors recognize this takes longer. This is more tough tech and traditional tech though the sales cycle and product cycle takes longer. There's more risk aversion across the field.
So that's why you've seen less than 3% of venture capital dedicated to a tech we think we need to improve that. There's a huge amount of opportunity for the public sector and private sector to be in partnership to accelerate this adoption cycle and get more venture capital into the sector. Thank you very much, Senator Boozman. Thank you, Madam Chair. And we do appreciate you all being here. It's interesting that a committee is really is for almost everything that we deal with is very bipartisan. The minority select witnesses, the majority select witnesses, but all of you all, again, could be witnesses for any of us.
Like I say, you're just trying to figure out answers, trying to come up with solutions in this very interesting field that has the promise of really making a huge difference to humanity in so many different ways. So thank you all for what you're doing, what you're working on. Dr. Hindman, in your testimony, you state John Deere, customers retain control over their data, including how it is collected, stored, processed, and shared. Can you talk a little more about how this works in practice? Including the process for producers who may wish to make changes in the way their data is being handled. Yeah, sure. Thanks for the question.
So today in our digital application, John Deere Operations Center customers, that's where they can control, modify, change the information that they have within, that's reflecting their particular farm operation. They get the opportunity to invite people within their organization to participate in that application And define the access rights for those people that might be other labor on the farm. It might be an agronomist that they're working with. It might be a third company that potentially is working with them on different aspects of their operation. But they get to determine what that information is, where it goes, and what it's used for.
And they can change that dynamically. They can change it on their mobile phone if they want. They can change it in the desktop application, and they can do that at any point in time. They also have the ability, if they so choose, to delete that information if they want to remove the record.
Generally, that's a pretty unlikely thing in our experience. But they have the capability and the opportunity to do that if they choose. Very good. Dr. Griffiths, what are the cybersecurity risk associated with integrating more technologies like AI into the agricultural sector? Why is it important to research and explore the transformative potential of AI in this area? Thank you for the question. Well, the risks are the same kinds of risks that you get with any AI application in artificial intelligence is, first of all, it's basically a set of algorithms or computer programs.
And it's data, whether it's test data or live data. And you have the opportunity to attack either. And hackers will and can attack either for whatever reasons they wish to do it. In the agricultural sector, we really need to start protecting the data. And that's part of the kinds of research that we're engaged in with industry and with South Dakota State University, for example. We've been looking at how data go from satellite to the cloud because we all put our data in the cloud these days.
And how can we ensure data confidentiality and integrity as the data move from one place to another? Because it's easy to attack those data. So we look at developing a secure, effective method, both for encrypting the data and then doing that at speed and cost without involving any delay in the processing of the data, which as you've just heard, it sort of operates some sort of milliseconds in time. It's very, very important for the data to get to the person who's making ultimate decisions from the AI.
So I think as AI evolves, we're seeing more and more technology being incorporated into the agricultural sector. And that increases what we call the threat landscape. And so there are more and more points at which you can begin to attack the system, the more points you have that, you know, you're only as strong as the weakest link as it were.
So I really think, well, I'll give you another example. When just before Russia invaded Ukraine, we saw pictures of tractors stuck in fields unable to move because the systems had been attacked. And that created an impetus for us to do some work on trying to protect farm vehicles from similar kinds of attack that could occur from unfriendly states. So we've looked at enhancing the security of these farming vehicles. Developing artificial intelligence enhanced what we call intrusion detection systems. People who are trying to get into the system to do damage, and these systems are designed to strengthen the cybersecurity of agricultural machinery.
So we're looking at this not just for today's technology, but we're looking five years out at the technology that will be in the fields to try and ensure that they're robust and cyber protected from the beginning. And I think it's important to remember that this entire supply chain, this food and security and clothing supply chain, it needs to be addressed at every single point in scale. So I often say, well, we talk a lot about protecting the data from sensors and from tractors and other vehicles, but we also have to protect the seeds. And the treatment of seeds and how the fungicides and herbicides are being measured to be put on seed, for seed production. Because we actually could see a significant latent effect on the actual development or non development of crops going on in the future. So I hate to be the negative person here.
I'm very excited about what artificial intelligence can do in the agricultural sector, but I feel like the child in the sixth sense. You know, I see cyber threats everywhere. And when I see a vulnerability, I think we have to find a way to attack it. So I just ask that you be fully aware that you can't separate any new technology these days from cybersecurity. The two have to go together.
And as technology evolves and as artificial intelligence evolves, it's multiple things. It's not a single thing we have to evolve, the cybersecurity that goes with it. Okay.
Thank you very much. Thank you, Madam Chair. Thank you very much, Senator Klobuchar. Thank you very much to both of you excited about the work that's going on.
This summer, I toured a number of firms in Minnesota that were making impressive use of precision agriculture. On one of the farms, the sprayer had been configured to only spray on plants that it correctly identified as weeds. I know you're seeing this all over in our agriculture communities. We know it helps reduce input costs better for conservation. That's why Senator Fisher and I are joining the Precision. I've introduced the Precision Agriculture Loan Act.
And I know Dr. Hindman, you mentioned the Precision Ag bill in your testimony which I appreciate. I guess I'd start with you, Mr. Krishnan. And in your testimony you discussed the important role that federal financing opportunities can have in helping scale these kinds of technologies. Could you talk a little more about that? I think as agriculture goes from data poor to data rich, everyone's talking about precision agriculture. So drones satellites, remote sensing, et cetera.
We also see a more data driven food and Ac system to offer new risk management and lending solutions. Particularly not only to adopt and de, risk the scaling of new technologies, but also create longer term decision making focused on soil health and sustainability. So I think that's a really important area for public policy to get involved in to give farmers not just the tools but the financing to increase yield breaker. And profit breaker. Exactly. I mean, Senator Thune and I
lead a different bill that would require the USDA to identify and collect and analyze the data. And so what you're saying is that it's helpful to both the data as well as the financing. I think both are really important and critical.
Okay. Dr. Earles, in your testimony, you talked about the work you are leading at UC Davis to develop AI enabled sensing systems that help producers manage their operations more precisely. Can you talk about how that can help with costs in the long term, it's investment in the short term, please. Thank you.
Yes, absolutely. So in terms of on the farm, you know, we think of the breakdowns in terms of cost. I think more of specialty crops because I'm coming from California, but I think this applies across many different types of crops where we have all of these inputs that farmers are facing. They're spread across a number of different activities, right? So AI has the potential to hit various activities such as whether it be fertilization, pest management, yield forecasting and prediction, and other types of irrigation, so on. I think these are cost saving measures that may be somewhere 5-15% on average in any given farmer's operation.
So one of the challenges in going from an idea to a product in agriculture for cost savings is finding those real value propositions. I think in precision agriculture, what we've been working on is really identifying what those are. That really depends on what crop type we're looking at. And I think this is a big challenge for AI going forward, is finding each one of those crop types pivot points, that they're willing to bring AI into adoption. You can't really do most of this without broadband, as we know, that is a big piece of this. To get broadband to every corner of our country.
We made sure that the bipartisan infrastructure law that many of us at this table supported, included a significant investment in broadband infrastructure, actually led that bill before it got included in the bipartisan infrastructure law. Dr. Griffiths, can you talk about the importance of broadband to making AI enabled a technologies? Farmers and how can we do more to solve the workforce shortage which is another issue that's plaguing us in rural? Absolutely. Thank you for that.
Yes. I'll start with the broadband question. Given the fact that we are able to potentially generate huge amounts of data, we have to send the data somewhere and we have to do it in a timely manner. And in order to do that, we need the broadband infrastructure to become, you know, ubiquitous across the entire country.
But especially in the farmlands and in the middle of the United States where it's not necessarily uniformly available at this time. That's been something actually that I've testified on before, and here it is again. We can't ignore that as an enabling infrastructure for the use of many of these additional and new and emerging technologies. Even in South Dakota, we're doing pretty well on broadband infrastructure on one side of the state and the other side, not so good.
And then we have bad lands and topologies that where it's very, very difficult to try and ensure that equal access to broadband technologies. But we're working on it and we'll continue to do so. The workforce issue is another issue that's plaguing us because, um, we really do not have sufficient numbers of people who are fully aware of the capabilities of artificial intelligence.
I classify the workforce needs into three areas. We have the need for additional experts who are actually going to help develop these AI related applications to agriculture and to other sectors. And we are in relatively short supply of those people.
We need to encourage more people to go into Stem. And I think that particular issue won't be solved with domestic personnel alone. We're going to have to look at legal immigration.
The second area though, we're going to have more people engaged in what we call the users. So we're going to two kinds of users of AI technologies. They're the producers of products and services who are using AI technologies to create their offerings.
And then we also have the end users of those products and services. So you could say the individual farmers or ranchers who actually need to use those products and services. And then we ultimately have the general public.
What should the general public know about artificial intelligence? So I think the key here is one, we do need more people moving into these fields. And secondly, we need to do more to educate the users and end users of the capabilities and the risks associated with these technologies. So that we can develop artificial applications, artificial intelligence applications responsibly that actually do what we want them to do. That carry the kinds of values the United States wishes to spread and continue to spread around the world. And minimize the risk that's associated with these technologies so that we can optimize the use.
Thank you. Thank you very much, Senator Ernst. Thank you very much, Madam Chair. And thanks to our witnesses for being here today as well. And and thank you, Senator Boozman for mentioning Steph Carlson.
We will miss Stephanie as she returns back to Iowa to be closer to her Iowa family friends and of course her great Iowa Army National Guard unit as well. So we really appreciate Steph and the great contributions that she has made to the committee and to my office. So thanks, Steph. So when we passed the Farm Bill of 2018, we thought it was a very significant accomplishment.
But here we are today, five years later, and we still have no Farm Bill. It actually expired on September 30. So I am encouraged to see that we do have a one year extension agreed to.
Contingent on the continuing resolution being passed. The lack of urgency and progress on this once every five year piece of legislation has been a real disservice to rural America. And I hope that we can see more farm in the Farm Bill as we move forward, and that we can get it done early this next year. We really need to work hard for our farmers and ranchers. I always make this point, many of you have heard it before, but I believe that food security is national security.
So we need to continue on and make sure that we get this over the finish line. The topic on today, it really does have so much potential for the future of Ag, whether it's identifying that specific type of insect, just as was pointed out. Thank you for that example. Or whether it is monitoring animal behavior in a hog barn. Ai has the ability to provide our farmers with new tools to help them navigate through very difficult decisions. So I am very proud to say that Iowa has been leading when it comes to AI.
And my alma mater, Iowa State University, is home to the AI Institute for Resilient Agriculture by partnering with Iowa farmers and companies like John Deere. And thank you very much, Dr. Man, for being here today. I am excited to see the future of agriculture together. They're focused on technology that makes Ag smarter, more profitable, and more sustainable to better meet the demands of our future generations. So thank you for engaging. So Dr. Hindman, I'll start with you.
And as you mentioned in your testimony, the utilization of AI in Ag requires a significant amount of user data which our Iowa farmers are collecting when they plant spray, fertilize, or harvest their crops. And my brother in law uses this technology as well. It's pretty exciting to hear him talk about the opportunities there.
So how do you ensure that the information and the privacy, which is something even my father, he has a good dose of Iowa farmers skepticism. How do you ensure that all of that privacy is protected for those farmers when they are utilizing your particular technology? Thank you. Center, it's a great question. You know, I think first I'll go back to they control who accesses their data within their account within the operation center. But before we even get to that stage, we think about we take this very seriously within Deere and we start with the principle of security by design. So making sure that as we're developing software, we're doing it with the seriousness about the security aspects of that software. First and foremost, making sure that it's secure by design.
That's sort of the bedrock of how we do software development, whether it's digital or embedded. Even with that, we still no doubt create opportunities for threat actors given the complicated threat surfaces that are involved to infiltrate that data. And so we then look towards external partners to help us do things like penetration testing, testing the systems to make sure that they are resilient to external people trying to get into those systems. We partner with White hat hackers. Hacker one would be an example where we do a bug bounty program.
We pay for ethical hackers to try to hack into our systems and expose vulnerabilities before they become public, so that we can remedy those and keep that data intact. That is, no, that's really good to know. And that will help some of these these older generations of farmers understand the new technology is safe to use and their data is safe. But once you've captured all of that data, then how is John Deere? Do you turn around and use that data to help our farmers? Yeah. It's used in a variety of ways for growers from,
you know, ways for us to look at how to improve the next product. Is the current product performing to the expectations of the customer or not. If not, what can we do to move the needle to make it perform better? We also utilize it for things like predicting when failures might occur in the equipment so that we can provide proactive services, proactive support for those failures, especially in those critical timing windows of planting and harvesting when, you know, every minute matters, machine downtime is a problem. And so we work to make sure that we can try to position both customers and dealers in the best position possible to be able to address those disruptions. That is really great, and the military has actually been using predictive maintenance and models for a very long time, which does save the federal government a lot of money.
And in this case, utilizing this technology on our farms may save our farmers a significant amount of money and downtime as well. So thank you very much. Thank you, Madam Chair. Absolutely. Senator Welch, thank you very much. Thank the witnesses. And I do want to thank Senator Stabenow, Senator Boozman, for the one year extension.
But time's a wasting and we've got a sketchy situation over there on the other side of the building. So we better get to work pretty quick to get that five year farm bill passed. There's a lot of opportunity with a, I'll start with you, but a real concern I have is for the viability of our smaller producers. I mean, we've got smaller farmers in Vermont, and a lot of times something will come up that has, it's an opportunity for bigger Ag, where you can spread the cost over time.
But for a lot of smaller producers, there's a lot of skepticism about whether they can get a return on investment. I saw that there was a Mckenzie study talking about North American farmers. 52% say the high cost and 40% cite the returns is the biggest challenges to adopting any farm system management opportunities. And a lot of our farmers have a lot more confidence in a fellow farmer or some of the folks from extension people that they've worked with over the years.
And they have confidence, understand the dynamics. And this is going to get to a point you're making about. The profitability over yield. But the big question I have about AI is with the entry costs. How can we do things that are going to help the smaller farmers get the benefit of it when they're just not going to be able to take that risk about the high upfront costs.
Can you address that, Dr. Earles? Absolutely, Judge. So I think there are two buckets that I would put what AI might help farmers do.
One is decision support. So this has typically been the role for small farmers of extension types of agents. They rely very heavily on extension, oftentimes for various sorts of advice. Right? And then the other one is mechanization and automation, which I think those two can often come with very different costs associated with them.
So I think in terms of AI and their impact on those two areas. On the extension side, I mentioned this idea of there aren't enough extension agents out there right now because really they are excellent in 100 different. Let me put a little on it too, because in addition to that, the smaller farmers, a lot of times will be wanting to do this focus on profitability and value added.
The data that goes into the AI algorithm is essentially generated by the larger operations. How is that going to make it more difficult for the smaller farms that have a different business model, in effect, to be able to take benefit of AI. How do we integrate that data into the systems? Yeah, I think there's an opportunity for using things people already have, like their phones.
And people are taking advantage of this because our phones are loaded with many different types of sensors. You know, we may not realize it, but there's probably 15 different types of sensors on there from cameras through accelerometers, et cetera. So I think there is an opportunity for smaller farmers and people are leveraging this opportunity to develop products around phones. Things let me talk to Thank you. Yes, it's helpful. I'll ask you,
Mr. Krishnan, I know your work. You focus on profitability per acre to better incentivize climate friendly practices. That's a big deal for our farmers. They're trying to do things that regenerate the soil. They are trying to do things that reduce the cost of a lot of these inputs so that their profitability is at the end of the day, obviously, what's important.
So talk to me about AI and how our farmers can focus on profitability as opposed to anything else really. Yeah, I think what I and better sensing allows you to do is focus on precision agriculture. So improve the resource efficiency of water, pesticide fertilizer management, and legal productivity. This should increase resource efficiency and profit per acre. I think the second thing that AI will allow us to do is sort of understand agriculture and the farm as a