AMPLIFY  VOL. 37, NO. 7
  

Global agriculture, the bedrock of our existence, faces unprecedented challenges from climate change, soil degradation, and water scarcity, affecting yields and threatening food security and sustainability. Escalating temperatures, unpredictable weather patterns, and environmental degradation endanger farmers’ livelihoods, our societal fabric, and the biosphere.

The agricultural sector’s contribution to the global economy has plummeted from 10% in the 1960s to below 5% in 2020, underscoring the crisis.¹ With the world’s population expected to reach 9.7 billion by 2050, sustainably increasing food production is paramount.²

Recent disruptions such as the Russia-Ukraine and Israel-Hamas wars, the pandemic, and ongoing environmental dangers highlight the fragility of our global food system. Without swift, decisive action, food prices could rise and supply chain disruptions could worsen.

Addressing in unison the three challenges of feeding a growing population, ensuring a decent livelihood for farmers, and safeguarding the environment is essential to sustainable progress. Because farmers constitute a large portion of our population, particularly in developing countries and rural areas, agricultural advances and productivity gains shouldn’t come at the expense of their livelihoods. And, of course, agricultural practices shouldn’t be detrimental to the environment.

The Role of AI

Food and agriculture were highlighted for the first time at the 2023 United Nations Climate Change Conference (COP 28). In this landmark event, more than 130 countries signed a “Declaration on Sustainable Agriculture, Resilient Food Systems, and Climate Action” to prioritize their food systems in their national strategies to combat climate change, signaling a global commitment to sustainable agriculture.³

Technological advances, particularly in AI, offer innovative and cost-effective solutions to boost agricultural productivity while mitigating economic and environmental risks. Technology can also bolster traceability in the agri-food chain, helping growers and distributors meet rising consumer demand for transparency about the origin of their food.

AI is a powerful tool with the potential to transform agriculture into a climate-resilient and sustainable sector. This article explores how AI, supported by the Internet of Things (IoT), cloud computing, and smartphones, can address the challenges posed by climate change and foster sustainable agricultural practices.

Agriculture 4.0

Like manufacturing and transport, agriculture has evolved over centuries. It is now in its fourth phase (Agriculture 4.0), which features improved crop yield, optimal resource allocation, and nutrient management. This phase’s technology-enhanced agriculture presents myriad ways to foster a journey toward climate-resilient, sustainable farming.⁴ Below are the key technologies driving modern agriculture:

• AI and predictive analytics. Machine learning (ML) and predictive analytics use data from satellite imagery, weather forecasts, and soil sensors to analyze crop health, potential yields, and the best times for planting, irrigating, and harvesting. They also predict pest invasions and disease outbreaks, enabling precise treatments that minimize crop damage and reduce chemical usage. Predictive analytics assist in soil management by optimizing irrigation and fertilization schedules. Generative AI (GenAI) platforms like ChatGPT can assist farmers by providing both general recommendations and specific information on their crops (see sidebar “GenAI in Agriculture”).

• Remote sensing and drones. High-resolution cameras and drone sensors capture detailed aerial images of crops and plants that AI can analyze to monitor crop health, identify pests and diseases, and evaluate crop maturity. This minimizes waste by helping farmers precisely apply water, fertilizer, and pesticides and increases yields.

• Robotics and automation. Autonomous robots can perform labor-intensive tasks like planting, weeding, and harvesting. They distinguish weeds from crops and selectively harvest ripe produce without causing damage. This precision reduces labor costs, saves time, and minimizes crop damage, resulting in healthier fields and increased yields. Robots can operate continuously, extending fieldwork hours.

• IoT and smart farming. IoT sensors gather real-time environmental data from soil moisture sensors, weather stations, and livestock health monitors. Analyzing this data helps optimize irrigation schedules, automate operations, and improve resource efficiency. Soil moisture sensors ensure optimal water usage, and livestock tracking devices offer early warnings about potential illnesses. IoT actuators can be remotely controlled to irrigate crops or apply pesticides when and where needed.

• Cloud computing. Cloud services host applications, data storage, and data and information-sharing facilities for the farming industry and the food supply chain. Farmers’ adoption of smartphones gives them access to these services.

Together, these technologies can transform agriculture into a data-driven, efficient, sustainable practice.

AI in Agriculture

AI and ML can empower agriculture by improving yield and resource use, minimizing waste, improving crop management, and ensuring that produce meets consumer demand.

Precision Agriculture

AI enables precise monitoring of soil health, water usage, and crop conditions through sensors and drones. John Deere’s See & Spray system (developed by Blue River Technology) uses computer vision to identify and target individual weeds with herbicides, increasing crop yield, reducing chemical use (by up to 90% in cotton fields⁵), and minimizing environmental impact while improving efficiency. John Deere reports a reduction in herbicide use by up to 77% compared to traditional broadcast spraying methods, resulting in significant cost savings and reduced environmental impact.⁶ This targeted approach also contributes to improved crop health and higher yields.

Automated Irrigation Systems

Companies like CropX use AI algorithms to analyze soil moisture data obtained via sensors and automatically control irrigation systems. CropX reports a 25%-50% reduction in water usage while maintaining or improving crop yields due to optimized irrigation scheduling.⁷ In addition, the CropX system is saving energy, fertilizers, and pesticides — reducing greenhouse gas (GHG) emissions and soil pollution on a large scale.

Crop Management, Yield Prediction & Optimization

AI-powered satellite imagery systems monitor crop growth, detect diseases, and recommend interventions. This proactive approach reduces losses and maximizes yield sustainably. For example, companies like Agrible and Descartes Labs use ML to analyze satellite data and provide farmers with insights into crop health, yield estimates, and potential issues like pests or diseases.⁸ Farmers can make data-driven decisions on irrigation, fertilization, and pest management.

A study published in Computational Intelligence and Neuroscience demonstrated an AI system’s capability to detect apple scab, a typical apple tree disease. Using a neural network trained on a data set of apple leaf images, the system achieved an impressive 95% accuracy in identifying disease presence.⁹

Robotics & Automation

AI-empowered robots could automate many labor-intensive agriculture activities, including preparing the soil, planting seeds, removing weeds, and harvesting. For example, autonomous weeding robots developed by FarmWise identify and remove weeds from crops without damaging plants. These robots offer benefits such as labor savings, reduced herbicide reliance, and improved weed control. Through better weed detection and removal, weeding robots reduce herbicide use, lowering costs and reducing environmental impact.¹⁰

Market Forecasting & Produce Supply Chain Management

AI empowers supply chains by predicting demand, optimizing logistics, and reducing food waste, resulting in sustainable consumption patterns and decreased GHGs. For example, AgShift uses AI to assess the quality and value of harvested crops, helping farmers make better pricing and market timing decisions. Through improved quality assessment, AgShift solutions reduce post-harvest loss and food waste, ensuring optimal use of harvested produce.¹¹

Climate Prediction & Adaptation

ML algorithms analyze climate data to predict weather patterns and optimize planting schedules. AI-driven models help farmers adapt cultivation practices to changing conditions, improving resilience against climate variability.

Data-Driven Decision-Making

AI-driven insights are transforming farming practices by enhancing data-driven decision-making. For example, Climate FieldView leverages AI to give farmers actionable insights from various data sources, offering personalized planting advice, optimized seeding rates, and guidance on the best times to water and apply fertilizers and pesticides. The program can predict how variables, such as weather changes or planting densities, might affect crop yields. This predictive capability lets farmers make more informed decisions, increasing crop productivity and reducing resource waste.

IBM’s Watson Decision Platform for Agriculture uses AI to analyze weather data, satellite images, and other inputs, providing farmers with actionable insights. The technology helps farmers precisely apply water, fertilizers, and pesticides, reducing overuse and runoff.

The World Economic Forum (WEF) is an excellent source for information on how data is transforming agriculture, including an article titled “How Is Agritech Helping to Optimize the Farming Sector?”¹² and one on how data-driven agritech services in Telangana, India, could add US $50-$70 billion to the agriculture sector by 2025.¹³

Huge Opportunities

Farmers can achieve significant efficiency gains and economic returns by leveraging AI-driven solutions while minimizing environmental impact. Indeed, AI presents unprecedented opportunities to build climate-resilient and sustainable agricultural practices. WEF elaborates on the benefits of AI in its article “Artificial Intelligence for Agricultural Innovation,”¹⁴ and specific use cases are described in the sidebar “AI in Action: Enhancing Agricultural Practices.”

However, concerted efforts by multiple stakeholders (governments, agritech companies, and farmers) will be needed to realize these opportunities. This may include subsidizing AI implementations and investing in agricultural AI R&D. To foster wider adoption, we recommend tailored education and training (e.g., Udemy courses) and the development of shared, trusted, open data platforms providing secure, role-based access to systems and stakeholders.

Adoption Challenges

AI’s true impact depends not on its vast potential but on large-scale adoption by farmers, smallholders, and large agricultural companies. Despite AI’s potential to foster sustainable agriculture, there are several major challenges to widespread adoption:

• Implementation costs. Many AI solutions require up-front investment in equipment and software, which can be prohibitive for small-scale farmers or those with limited financial resources. The initial cost of adoption may outweigh the perceived benefits, especially if the ROI may not be realized for several years.

• Complexity and integration challenges. AI technologies often require integration with existing farming practices, equipment, and data systems. The complexity of implementation and potential disruptions to established workflows can deter farmers from adopting these solutions.

• Data privacy and security concerns. AI applications in agriculture rely heavily on data collection and analysis, including sensitive information about crop yields, soil health, and farming practices. Farmers may have concerns about data privacy, ownership, and security when using AI-powered platforms and services.

• Limited access and unreliable connectivity. Many areas where agriculture is prevalent have limited access to reliable Internet connectivity and infrastructure essential for real-time data collection and analysis.

• Tailoring to local contexts. AI solutions must usually be tailored to local agricultural contexts, including specific crops, soil types, climate conditions, and farming practices. One-size-fits-all solutions are unlikely to effectively address the diverse needs of farmers.

In addition, Western technology and business models often don’t apply in other geographies due to differences in scale. Western farms are typically larger than 1,000 acres, while farmers in India and many developing countries usually manage between two and 10 acres, often in different locations. Technologies designed for large-scale farms are not practical for smallholders.

To address these challenges, agritechs must focus on effective distribution and adoption rights. Village-level entrepreneurs (VLEs) play a crucial role in this process. Connecting VLEs with private capital is essential to driving social development and change. Furthermore, technology must be localized to meet the specific needs of these regions:

• Lack of awareness and education. Farmers may not be fully aware of the capabilities and benefits of AI technologies in agriculture. They may lack education and training on effectively integrating these technologies into their farming practices.

• Resistance to change. Farming is a traditional industry in which practices are often passed down through generations. Resistance to change and reluctance to adopt new technologies can be cultural and/or fear-based.

• Risk aversion. Farmers may be risk averse when adopting new technologies that could impact their livelihoods. The potential failure or disruption caused by adopting AI solutions may outweigh the perceived benefits for some farmers.

• Regulatory and policy barriers. Regulatory frameworks and policies related to AI adoption in agriculture vary across regions and countries. Uncertainty about compliance and regulations may create barriers to adoption.

A recent study in Environmental Science & Policy offers farmers’ perspectives on options for and barriers to implementing climate-resilient agriculture and implications for climate adaptation policy.¹⁵

Recommendations

To promote the widespread adoption of AI in agriculture, stakeholders such as governments, technology providers, research institutions, and agricultural organizations must work together to address challenges related to cost, education, infrastructure, regulation, and cultural acceptance.

The WEF initiative “AI for Agriculture Innovation (AI4AI)” is designed to advance the global agenda of digital agriculture by scaling up emerging technologies and encouraging public-private partnerships. The initiative aims to impact 1 million farmers globally by 2027. Currently, more than 50 organizations are involved.¹⁶

AI4AI’s resources section provides templates, playbooks, and learning resources for national and sub-national governments to structure their digital agriculture initiative. It also presents impact stories on AI innovation and adoption in India.¹⁷ Individuals and organizations are invited to join the forum and help shape a better, more sustainable future.

AI4AI’s Saagu Baagu initiative (the name means “agricultural advancement” in Telugu) has significantly improved the livelihoods of 7,000 chili farmers in India’s Telangana district by enhancing yields and incomes through advanced agritech and data management. Supported by the Bill & Melinda Gates Foundation and implemented by Digital Green, the project resulted in a 21% increase in chili yields, reduced pesticide and fertilizer use (by 9% and 5%, respectively), and increased product prices by 8%, effectively doubling farmers’ incomes.¹⁸ Encouraged by this success, Telangana’s government plans to expand the project to impact 500,000 farmers across five crops and 10 districts, demonstrating AI’s transformative potential.

To fully harness AI’s potential in building climate-resilient agriculture, concerted efforts by multiple stakeholders are needed:

• Governments and the agritech industry must invest in AI R&D for agriculture. They should also drive and support technology adoption among farmers.

• AI and agriculture researchers must collaborate to devise and develop viable AI solutions tailored to diverse agricultural contexts.

• Research institutions and significant stakeholders should focus on interdisciplinary studies to address AI’s socioeconomic and ethical implications in agriculture.

• Governments should promote and subsidize implementations to foster AI in agriculture to make it climate-resilient and eco-friendly.

Here are a few recommendations to address the challenges in adopting AI and other technologies in agriculture:

• Promote education and training. Educate farmers and stakeholders about AI technologies, including their benefits, risks, and limitations.

• Develop open data platforms. Encourage the creation of open data platforms that facilitate the sharing of agricultural data and AI models for collective learning and innovation.

• Prioritize ethical AI. Ensure AI technologies adhere to ethical guidelines, including respecting data privacy, fairness, and inclusivity.

• Choose affordable alternatives. To reduce the cost of cloud computing services, farmers should consider avoiding large providers like Google, Microsoft, and Amazon. For example, Flexisaf, an African tech company, found a solution to its problem (the high cost of Google’s Workspace) in Zoho, an Indian company that offered similar products at a lower price.¹⁹

In a recent article, WEF used lessons from a real-world project in India to show how AI can transform agriculture.²⁰ The article recommends adopting a public-private partnership framework. Other WEF recommendations include central and state government support and incentives, including incubators, start-up funding, government-backed venture capital, tax holidays, and exemptions.

Agritechs must focus on effective distribution and adoption rights to address the high cost of mechanizing agriculture, a barrier to making small rural farms more efficient and profitable. VLEs play a crucial role in this process, and connecting VLEs with private capital is a pathway to social development and change.²¹ Technology must also be localized to meet the specific needs of each region.

Late last year, WEF identified seven key innovations needed to transform food and agriculture:²²

1. Accurate, accessible weather forecasts. With extreme weather threatening crops and challenging farmers’ adaptability, accurate weather forecasts are essential. Farmers need to anticipate short- and long-term conditions to make strategic decisions about planting, irrigating, fertilizing, and harvesting. For example, accurate state-level forecasts of seasonal monsoon rainfall could help Indian farmers optimize sowing and planting times, potentially providing an estimated $3 billion in benefits over five years at a cost of around $5 million.²³

2. Microbial fertilizers. These fertilizers use bacteria to enhance the absorption of essential nutrients by plants and soil, reducing the need for nitrogen fertilizers (a significant source of GHGs). GenAI could assist in developing new types of fertilizers.

3. Reducing methane emissions from livestock. Livestock accounts for roughly two-thirds of agriculture’s GHGs, so implementing measures to reduce methane emissions is crucial.

4. Rainwater harvesting. Helping farmers and communities implement effective rainwater harvesting techniques can significantly improve water availability and management.

5. Lowering costs of digital agriculture. Making digital agriculture technologies more affordable can help farmers optimize irrigation, fertilizers, and pesticides.

6. Encouraging alternative proteins. Promoting the production of alternative proteins can reduce the demand for livestock, lowering GHGs.

7. Providing insurance and social protection. Insurance and social protections that enhance people’s well-being by reducing poverty and vulnerability through policies and programs can help farmers recover from extreme weather events and build resilience.

Conclusion

AI presents unprecedented opportunities to build climate-resilient and sustainable agricultural practices. Integrating AI into agriculture systems can enhance productivity, preserve resources, and mitigate climate risks. However, realizing this vision requires collaborative action and ethical stewardship to ensure that AI benefits all stakeholders and contributes to a more sustainable future.

References

¹ “Share of GDP from Agriculture.” Our World in Data, 24 May 2024.

² “Global Population Is Growing.” EU Competence Centre on Foresight, 21 December 2022. 

³ “Emirates Declaration on Sustainable Agriculture, Resilient Food Systems, and Climate Action.” Council of the European Union, 16 November 2023. 

⁴ B.P., Sreekanta Guptha, et al. “Agriculture 4.0 — A Journey Towards Sustainable Farming.” Infosys, 18 May 2022. 

⁵ “Pesticides.” Cotton Today, accessed July 2024.

⁶ “John Deere Launches See & Spray Ultimate Technology.” John Deere, 7 March 2022.

⁷ CropX website, 2024.

⁸ The Descartes Labs Platform, accessed July 2024.

⁹ Wang, Guan, Yu Sun, and Jianxin Wang. “Automatic Image-Based Plant Disease Severity Estimation Using Deep Learning.” Computational Intelligence and Neuroscience, Vol. 2017, No. 1, January 2017.

¹⁰ Bernier, Catherine. “Weeding Robots: Redefining Sustainability in Agriculture.” HowToRobot, 17 July 2023.

¹¹ AgShift website, 2024.

¹² Goel, Arushi, and Sowmya Komaravolu. “How Is Agritech Helping to Optimize the Farming Sector?” World Economic Forum (WEF), 31 October 2023.

¹³ Neo, Gim Huay, and K.T. Rama Rao. “How to Make Digital Transformation of Agriculture Work. Lessons from Telangana.” WEF, 15 June 2023.

¹⁴ “Artificial Intelligence for Agriculture Innovation.” WEF, March 2021.

¹⁵ Kundu, Shilpi, Edward A. Morgan, and James C.R. Smart. “Farmers Perspectives on Options for and Barriers to Implementing Climate Resilient Agriculture and Implications for Climate Adaptation Policy.” Environmental Science & Policy, Vol. 151, January 2024. 

¹⁶ “AI for Agriculture Innovation (AI4AI).” WEF, accessed July 2024.

¹⁷ “Impact Stories.” WEF, accessed July 2024.

¹⁸ M.S.V., Janakiram. “How Indian Farmers Are Using AI to Increase Crop Yield.” Forbes, 1 February 2024.

¹⁹ Dosunmu, Damilare, and Ananya Bhattacharya. “African Tech Companies Are Ditching Google for a Small Indian Competitor.” Rest of World, 22 April 2024.

²⁰ Neo and Rao (see 13).

²¹ Limaye, Pushkar, Devansh Pathak, and Sandeep K. Sinha. “How Rural Entrepreneurs Are Driving Agritech Adoption.” WEF, 4 April 2023.

²² Winters, Paul. “These Are the Seven Key Innovations Needed to Transform Food and Agriculture.” WEF, 7 December 2023.

²³ Winters (see 22).