Summary
- The use of AI in food animal farming requires consideration of data ownership and quality, as well as a proactive approach to data management.
- Building a comprehensive database of farm-specific data is crucial for improving decision support and intelligence on the farm.
- Data quality is the most significant bottleneck to the appropriate use of AI on farms, and a data quality mindset is necessary for improving it.
- Increasing digital literacy levels within the supply chain and service sector is crucial for increasing the impact of sustainable production and improving food animal welfare.
- A customized approach to AI and support tools are necessary for providing independent business intelligence and decision support to each farm, with greater ROI.
As we observe the clash of the ginormous tech companies that we all commonly know and see in our day-to-day lives, betting their future prosperity on AI as the catalyst to the next industrial revolution, food animal producers may ask how will the “new AI” help, what will it change, what will it improve and how should one prepare to evaluate the benefits and risks to their farm business?
As we are still in the first generation of AgriTech tools and identifying the existing limitations of “black box” algorithms, data bottlenecks and silos, data ownership challenges, and low digital literacy rates that impede on-farm and service suppliers with challenges associated with improving animal welfare and sustainability. So, before we go barreling towards AI use on-farm, we have a couple of things to consider ahead of our “AI Hype Cycle” in food-animal agriculture. These are:
- Data ownership, where farmers own their data pre- and post-algorithmic processing.
- AI and algorithmic processing for improving decision support and intelligence require quality data that can be collected, aggregated, integrated, assessed, managed, and stored.
Creating a comprehensive database at the farm level is crucial in ensuring the quality and reliability of the collected data. This database will act as a central hub that aggregates the unique data produced on each farm, making it easier to assess and interpret. Having a clear understanding of the different data sources can quickly identify and resolve any issues or errors. It is important to determine whether the problem is a result of a farm-related issue or a technical issue, which is why data ownership is so crucial in the context of AI and food animal management. The cost of mistakes and failed models is too high to ignore, making it imperative that we take a proactive and strategic approach to data ownership, quality, and management.
Farm data has value; third parties can use it for the development of new tools and insights that can inevitably be sold back to farmers, so understanding “Where does your farm data go” is critical to maintaining the value of your data. Aggregating farm data enables farm decision-makers to choose what AI and ML work for the decision support and intelligence needed on the farm, with appropriate governance and user license to third parties and off-farm users to track where your data is going and how it is being used and eliminating vendor lock-in. Farmers are a central part of the next generation of AgriTech and should be appropriately supported and remunerated for their supply of valuable data, their reward for harvesting quality data.
Farm data has value; third parties can use it for the development of new tools and insights that can inevitably be sold back to farmers, so understanding “Where does your farm data go” is critical to maintaining the value of your data. Aggregating farm data enables farm decision-makers to choose what AI and ML work for the decision support and intelligence needed on the farm, with appropriate governance and user license to third parties and off-farm users to track where your data is going and how it is being used and eliminating vendor lock-in. Farmers are a central part of the next generation of AgriTech and should be appropriately supported and remunerated for their supply of valuable data, their reward for harvesting quality data.
As we come out of the first generation of AgriTech tools in food animal farming, towards the next, we will see a greater emphasis on quality versus just “big” data. Today like the milk in our tanks is to data, quality is imperative for accurate decision support. As the saying goes, Garbage In = Garbage Out (GIGO) with data collection. This is singularly our major bottleneck to the appropriate use of AI on farms, as getting highly accurate AI model outputs relies on one thing—high-quality data as the input. To combat this on-farm, here are eight proactive steps to begin improving data quality:
- Build and foster a data quality mindset in your organization.
- Enlist data quality champions and data stewards within your team.
- Invest in people with the right skills and through additional training as necessary.
- Decide what metrics to use to measure your data quality.
- Identify and break down data silos on-farm by consolidating data into a farm-owned data ecosystem.
- Begin data quality assessment efforts when data is created and consolidated.
- Develop a process to regularly assess and maintain data quality and automate where necessary.
- Deploy tools to help continuously monitor and manage data quality.
A data quality mindset with the farm team is required. It is the first major step in digital literacy on the farm. Understanding why a single data point is important and what its value is to decision support on the farm should help the team’s “light bulb” to go off. Once this happens, it should then become easier to understand why aggregation and consolidation of farm-specific data, automating processes, and building a farm data ecosystem are important and valuable.
This means not just improving farm digital literacy levels but also within the supply chain and service sector. Food animal agriculture does not need more farmers but does need more AgriTech-enabled farm team members; this is part of the interdisciplinary skills training that can attract and retain talent, give opportunities and create cross-industry creativity and innovation both on and off the farm. In a recent OECD report, the share of general services for agriculture (such as innovation, biosecurity, and infrastructure) is small and has declined to 13% of support in 2019-21, down from 16% two decades earlier. Reversing this decline in the farm service sector is vital for increased impact on sustainable production and reducing greenhouse gas (GHG) while improving food animal welfare at the farm level.
We also need to consider that no two farms are the same; the cookie-cutter approach to AI and support tools do not enable appropriate independent business intelligence and decision support, but this is one of the greatest opportunities within food-animal agriculture, generating a digital services multiplier effect. The collection of valuable and quality data from each farm allows greater application of ML and AI to the specific needs of the farm more accurately, to the style and region, with greater ROI, that aids in reaching our business and industry targets, including improved sustainability and food animal welfare.
But, before all the AI hype in food animal agriculture, we have our eight steps to remember; we’ve got to effectively master and manage ML operations before deploying AI, accept that farmers own their data pre- and post-algorithmic processing through the data infrastructure that is needed to manage, aggregate, and store data independent of third parties. Once we have a quality data ecosystem comprised of the highest priority different sources of valuable data that is farm specific, we can then start applying AI that is of specific value to the business.
With a sustained pipeline of quality data from the farm, it will be incredible to see the applications that can be applied, like a farm version of ChatGPT or digital twinning and related technologies. Intelligently growing data points of value that feed into the data ecosystem and better understanding how that broadens the applications of quality AI to the data ecosystem is also an agile mindset, which is also what we want to foster within food animal AgriTech at the farm level.
By adopting these fundamental measures in the early stages of AI integration in food-animal agriculture, we can lay the foundation for promoting food-animal welfare and sustainability as a norm within the industry. This not only supports the ethical treatment of animals but also helps to promote environmentally sustainable practices. With this new data and information from the farm level, the story of food animal welfare and sustainability can be traced up the supply chain and communicated effectively to the consumer, empowering them to make informed choices about the food they purchase. By collaborating with farms at the farm level to prioritize animal welfare and sustainability, the food industry can make significant progress toward a more responsible, sustainable future and new value for farmers.
This means not just improving farm digital literacy levels but also within the supply chain and service sector. Food animal agriculture does not need more farmers but does need more AgriTech-enabled farm team members; this is part of the interdisciplinary skills training that can attract and retain talent, give opportunities and create cross-industry creativity and innovation both on and off the farm. In a recent OECD report, the share of general services for agriculture (such as innovation, biosecurity, and infrastructure) is small and has declined to 13% of support in 2019-21, down from 16% two decades earlier. Reversing this decline in the farm service sector is vital for increased impact on sustainable production and reducing greenhouse gas (GHG) while improving food animal welfare at the farm level.
We also need to consider that no two farms are the same; the cookie-cutter approach to AI and support tools do not enable appropriate independent business intelligence and decision support, but this is one of the greatest opportunities within food-animal agriculture, generating a digital services multiplier effect. The collection of valuable and quality data from each farm allows greater application of ML and AI to the specific needs of the farm more accurately, to the style and region, with greater ROI, that aids in reaching our business and industry targets, including improved sustainability and food animal welfare.
But, before all the AI hype in food animal agriculture, we have our eight steps to remember; we’ve got to effectively master and manage ML operations before deploying AI, accept that farmers own their data pre- and post-algorithmic processing through the data infrastructure that is needed to manage, aggregate, and store data independent of third parties. Once we have a quality data ecosystem comprised of the highest priority different sources of valuable data that is farm specific, we can then start applying AI that is of specific value to the business.
With a sustained pipeline of quality data from the farm, it will be incredible to see the applications that can be applied, like a farm version of ChatGPT or digital twinning and related technologies. Intelligently growing data points of value that feed into the data ecosystem and better understanding how that broadens the applications of quality AI to the data ecosystem is also an agile mindset, which is also what we want to foster within food animal AgriTech at the farm level.
By adopting these fundamental measures in the early stages of AI integration in food-animal agriculture, we can lay the foundation for promoting food-animal welfare and sustainability as a norm within the industry. This not only supports the ethical treatment of animals but also helps to promote environmentally sustainable practices. With this new data and information from the farm level, the story of food animal welfare and sustainability can be traced up the supply chain and communicated effectively to the consumer, empowering them to make informed choices about the food they purchase. By collaborating with farms at the farm level to prioritize animal welfare and sustainability, the food industry can make significant progress toward a more responsible, sustainable future and new value for farmers.
Reference Items
- 8 Proactive Steps to Improve Data Quality (techtarget.com)
- The importance of improving data quality at source | Computer Weekly
- Data Quality: The Real Bottleneck In AI Adoption (forbes.com)
- Automated detection of poor-quality data: case studies in healthcare | Scientific Reports (nature.com)
- Sensors | Free Full-Text | Machine Learning in Agriculture: A Review (mdpi.com)
- McKinsey’s latest AI survey shows continuing growth in usage and reveals critical best practices. | McKinsey & Company
- Innovation & other general services can support long-term goals – OECD
- The Data Dilemma: Four Common Barriers To AI Success (forbes.com)