2019 AI in Agriculture: A Snapshot
-
Agricultural AI is steadily advancing, transitioning from conceptualization to practical field applications. This article reviews the applications of AI in agriculture and provides insights into its future prospects.
Over the past year, 'AI+' has permeated various sectors of Chinese industry, from industrial quality inspection to smart cities, revealing an increasingly clear vision of the Fourth Industrial Revolution.
However, if we shift our focus beyond urban peripheries to the vast expanse of China's rural landscape, can AI take root in agriculture, propelling this millennia-old primary industry toward higher levels of industrialization?
The answer is yes.
In 2019, we witnessed AI technologies such as computer vision, deep learning, edge computing, and intelligent robots being employed to enhance agricultural productivity. From highly digitized pig and poultry farms to intelligent sorting and harvesting robots, AI has begun delivering tangible value by transforming the agricultural supply chain through cutting-edge technology.
Yet, we also recognize that for agriculture to fully harness AI's potential, it must first undergo the industrialization processes typical of secondary industries and the socialized workflows of tertiary industries. Without this progressive industrial foundation, the promising vision of 'AI + agriculture' may dissipate like a passing shower, failing to penetrate deep into the roots.
How can the nourishing rain of intelligence reach across 9.6 million square kilometers of land? In 2019, agricultural AI embarked on a mission to sow the seeds of a smarter future.
BIS Research recently released a report titled 'Global Artificial Intelligence (AI) in Agriculture Market Analysis and Forecast, 2019–2024.' The latest market intelligence indicates that the agricultural AI market was valued at an estimated $578 million in 2019 and is projected to grow at a compound annual growth rate (CAGR) of 28.38%, reaching $2.0157 billion by 2024. While demand-driven agricultural intelligence seeks to seize this era's opportunities, the path is far from straightforward.
The core challenge lies in the stark differences between agriculture as a primary industry and the more modernized secondary and tertiary industries.
Thus, in this article, we might approach the topic differently by first revisiting the origin of the problem—what pressing issues does today's agricultural supply chain hope to solve through AI's complex algorithms?
In recent years, the transformation from labor-intensive to industry-intensive practices has become the dominant theme in Chinese agriculture. Key drivers of this shift include low per-unit crop yields, the prevalence of small-scale family farming, and the declining interest of younger generations in farming as a profession. Amidst policy trends like environmentalism and industrial consolidation, replacing manual labor with intelligent machinery has emerged as a defining theme for 2019 and beyond.
Addressing these challenges inevitably raises a second issue—the high technical barriers of AI in agriculture. China's long-standing small-scale farming economy and policy-driven technology dissemination model make initial costs, safety concerns, and other factors obstacles to intelligent and large-scale agricultural management.
Although robotics and smart algorithms have simplified certain processes, smallholder farmers still account for over 80% of China's agricultural workforce, with an average education level below 7.5 years. This lack of technical expertise hampers the translation of AI advancements into real-world productivity.
Even with government subsidies, financial insurance, and other mechanisms, the high costs of standardizing and modernizing production will inevitably affect agricultural product prices in the short term. The recent surge in pork prices, along with increases in beef, lamb, and egg prices—and even temporary spikes in fruit prices—highlights the disconnect between production and consumption. This 'low prices hurt farmers, high prices hurt consumers' dilemma underscores the need to integrate agriculture into urban digital economies.
In 2019, AI applications in agriculture moved beyond showcase projects like 'AI pig farming' or isolated automation upgrades, instead extending deeper into the soil with a more comprehensive approach.
Three key trends emerged in 2019: First, AI applications in agriculture became more integrated and holistic.
This was particularly evident in production. If 2017–2018 marked the experimental phase of AI in agriculture, 2019 saw the emergence of deeply cross-disciplinary solutions combining AI and farming.
From deploying edge-computing sensors to applying vision perception, language reading, logical reasoning algorithms, and human-machine collaboration, AI in agriculture transitioned from isolated tasks to comprehensive transformations.
For instance, a Yunnan-based agricultural equipment manufacturer developed an intelligent edge platform that, combined with cloud-based data training, delivers vertical algorithm models to production equipment, enabling self-adjusting operational parameters. This AI + IoT solution achieves production quality comparable to that of mid-level skilled workers.
Additionally, AI algorithms in agriculture improved in accuracy and practicality. In 2019, machine recognition moved beyond lab settings, overcoming regional and crop-specific challenges to deliver higher precision and reliability for farmers.
One company partnered with an AI firm to install specialized brackets and cameras in greenhouses, automatically assessing crop ripeness, optimizing growth conditions, and identifying pest infestations. This enabled robotic sorting and reduced losses from unforeseen issues.
In Hainan, hundreds of farms adopted smart management systems for IoT-based monitoring. In Xinjiang, remote-controlled cotton-picking machines harvested 60 acres per hour—1,000 times faster than manual labor. In Inner Mongolia, a herder equipped over 300 yaks with 5G devices, paving the way for 'remote herding.' These examples show how AI is reducing costs not just on paper, but in practice.
Another notable shift was the evolution of agricultural technology dissemination from government-led models to tripartite collaborations involving government, enterprises, and academia. Traditional hierarchical extension networks are giving way to partnerships between tech firms and agricultural giants, fostering tighter integration of policy, innovation, and technology.
For example, a financial institution used AI models to analyze multi-dimensional farmer data, enabling rapid credit scoring for pig farmers. This increased financing opportunities while lowering costs, helping address cyclical price fluctuations ('pig cycles').
Such initiatives meet farmers' needs more effectively, translating innovations into productivity. They also alleviate fiscal pressures by supplementing public funding with private investment, reducing R&D and dissemination costs for tech firms, and expanding the reach of new technologies.
Overall, these mature, multifaceted, and wide-ranging applications are expected to accelerate AI adoption in agriculture over the coming years.
Giving agricultural AI an 'A' for its 2019 performance is only natural. However, the dawn of intelligent, networked transformation also means AI faces more challenges on the path to precision agriculture.
For instance, the coverage of agricultural AIoT networks needs expansion. Many innovations rely on village-level digital infrastructure, particularly IoT and cloud computing, to enable real-time data processing and decision-making.
According to statistics, the internet penetration rate in rural areas of China is 36.5%, only half of that in urban areas. For AI to take root across the country's 9.6 million square kilometers, the first challenge to address is the 'data scarcity syndrome,' which likely depends on the comprehensive rollout of the next-generation internet and IoT deployments.
Meanwhile, Chinese tech companies' involvement in agriculture currently remains at the stage of infrastructure transformation and algorithm empowerment. In the future, collecting and open-sourcing high-quality, valuable agricultural datasets may prove to be the most effective remedy for accelerating agricultural AI progress.
Additionally, there is still a shortage of specialized chips for smart agricultural equipment. Most current AI applications are built on general-purpose chips. However, unlike highly standardized factory or urban environments, agricultural smart devices face complex production scenarios and ever-changing environmental conditions. Such chips are prone to damage in the harsh conditions of fields, affecting the reliability of smart agricultural machinery. Currently, the influence of agricultural demands on driving the semiconductor industry chain remains insufficient.
In terms of services, addressing issues such as the lack of willingness and capability among some family farmers to adopt AI, agricultural financial credit risks, and the branding of agricultural products requires relevant authorities or social enterprises to use AI to establish vertical industry prediction models. These models can guide and help agricultural producers dynamically adjust their production activities. Determining how to support service providers offering such B2B and B2C solutions has become a multiple-choice question awaiting answers from agricultural AI.
Overall, these challenges are both historical issues from 2019 and gifts from the future.
In 2017, the 'New Generation Artificial Intelligence Development Plan' proposed measures to promote the intelligent upgrading of agriculture, establish typical agricultural big data intelligent decision-making analysis systems, and carry out integrated application demonstrations such as smart farms, intelligent plant factories, and smart processing workshops for agricultural products.
Today, AI has blossomed across fields, overcoming the inherent complexity of agriculture and various technical implementation constraints, nurturing numerous AI oases and giving rise to many excellent solutions.
From spring sowing to autumn harvest and winter storage, what AI has written this year is precisely the hope and deep affection for this land.