[Special Interview] Interview with Son Jeong-ik, Director of the Smart Farm R&D Project Group

 

Facing unprecedented survival challenges of the climate crisis and a declining agricultural population, the breakthrough for our agriculture lies in ‘technology (Tech).’ The Smart Farm R&D Project Group (hereinafter, the Project Group), as the control tower of a multi-ministerial package project that broke down barriers between ministries, has led a major transformation of South Korea’s smart farm R&D toward field verification and industrialization. Through an in-depth interview with Son Jeong-ik, Director of the Smart Farm R&D Project Group (emeritus professor at Seoul National University), this paper presents the technological peak reached by K-smart farms, their global competitiveness, and a blueprint for future agriculture in which humans and artificial intelligence (AI) coexist.

What has been the most significant achievement of the Smart Farm R&D Project Group, which was approved in 2020 and began full-scale activities in 2021?
“It is that we transformed smart farm R&D, which had remained at the level of individual technology development, into an integrated system centered on field verification and industrialization. Through the first-phase project, we secured core technologies such as the ‘high-temperature, high-humidity smart greenhouse package for export’ (winner of an industrial award) and ‘selective light-transmitting solar cells’ (selected among the Top 100 Outstanding National R&D Achievements in 2024), and achieved about KRW 1.8 billion in on-site sales with ‘livestock labor-saving robots.’
Currently, in the second phase, we are developing next-generation technologies such as digital twins, generative AI-based crop and livestock management, and agricultural work robots. In particular, building the ‘Korean Smart Farm Integrated Platform (KoSToP),’ which gathers all these smart farm research outcomes (big data, AI models, etc.) in one place for use in various projects, industrialization research, and new R&D, is the greatest foundation for our agriculture to leap forward on an AI basis.”

What differentiates K-Farm, the Korean-style smart farm model, in the global market, and what efforts is the Project Group making for this?
“The core competitiveness is the ‘K-Farm package model,’ which does not simply sell technology but provides facilities, equipment, data analysis, and remote management in an integrated package tailored to local climates. The Middle East and Southeast Asia are hot and humid, so local optimization is essential.
Through the project ‘Demonstration of the K-model smart poultry house for export,’ the Project Group is building a cooperation system with local governments, universities, and companies in Malaysia while analyzing and verifying the environment. The success or failure of overseas expansion lies not in ‘technology export’ but in ‘thorough localization,’ and we will secure sustainable competitiveness by supplying even post-operation training for operators as part of the package.”

What is the ideal model of ‘partially autonomous collaboration between humans and AI’ that you emphasized?
“Current smart farm technology has entered the AI-based autonomous stage (3–4.0) and is being partially applied in the field. Although 100% full autonomy without any human intervention may be technically possible, unless it is for a special purpose, economic feasibility issues arise. Therefore, an ‘appropriate level of autonomy’ suited to each farm’s conditions is desirable. The most ideal coexistence model is one in which autonomous robots take over not only difficult physical labor but also tasks that cause mental stress to farmers, allowing people to farm with enjoyment.”

How are ‘digital twin’ and ‘synthetic data’ technologies actually improving research efficiency?
“To overcome cost and environmental limitations, we are actively utilizing virtual spaces and generative AI. A representative example is the project ‘Commercialization of greenhouse production systems using digital twins,’ through which we implemented a virtual environment identical to a real greenhouse. Based on growth data spanning 60 cropping cycles, we repeatedly verify growth and quality according to changes in temperature and light intensity with an AI simulator before actual cultivation. In addition, we are developing a ‘generative AI-based decision-making platform’ that integrates environment, pests and diseases, and cultivation manuals for tomatoes, cucumbers, and paprika, dramatically reducing trial and error in research.”

What efforts is the Project Group making to address ‘data standardization,’ a major difficulty for companies and farms?
“If data structures differ, integration and AI training require enormous costs. To solve this, we introduced DMP (Data Management Plan) from the research initiation stage and are managing the entire lifecycle according to guidelines. In addition, we standardized information items based on the standard codes of the Ministry of Agriculture, Food and Rural Affairs to maximize interoperability with other platforms.
In particular, starting this year, we are advancing the platform based on MCP (Model Context Protocol), a global open protocol released by Anthropic, so that generative AI can access data in a standardized way. Through this, an ecosystem will open in which AI models from companies and research institutions can safely utilize KoSToP’s high-quality data in real time.”

Even if advanced technology is introduced, it is useless if field accessibility is low. What are the alternatives for technology diffusion and education?
“As important as technology development is ‘cultivating people who can use it.’ From 2021 to 2024, the Project Group has operated demand-tailored education programs centered on agricultural AI, big data, and automation systems to foster professional talent. Recently, we have been promoting the production of ‘specialized technical books’ that comprehensively compile our research achievements, core technical principles, and actual application cases. These technical books are scheduled to be published by next year in a total of 20 volumes—10 in livestock and 10 in horticulture—and if they are used as practical guidelines for industry, academia, research sites, and students, the achievements of laboratories will quickly spread to farms and companies.”
 

From the perspective of farms, economic viability (ROI) is key. Are there alternatives to reduce investment and maintenance costs?
“According to the Ministry of Agriculture, Food and Rural Affairs’ First Smart Agriculture Promotion Act, the goal is to raise smart farm area to 35% of the country’s total greenhouse area by 2029, and the ministry announced that it stood at 16% as of the end of 2024. Also, 90% of all greenhouse types are for vegetable cultivation, and 90% of those vegetable greenhouses are ‘single-span plastic greenhouses,’ so if highly advanced smart farm technologies are投入, ROI does not emerge. To expand smart farm dissemination, smart farm R&D must prove not only performance improvement but also economic viability (ROI). For smart farm dissemination policy to be effective, a low-cost Korean-style dissemination system is essential. To this end, the Project Group is currently conducting a precise economic analysis of the K-Farm model (B/C ratio, investment payback period, break-even point, etc.). Based on these data, we plan to verify which crop and technology combinations increase profitability and reduce maintenance burdens, and present a customized mid- to long-term dissemination strategy for farms to the government.”

What business models and policy support are needed for urban vertical farms to become self-sustaining?
“For urban vertical farms, the priority is to establish independent business models with clear purposes, such as supplying fresh vegetables tailored to urban consumption patterns, regular deliveries to hotels and restaurants, small-scale production of premium crops, and subscription-based salads. On the policy side, support is needed for remodeling costs of vacant urban commercial spaces, rent, and energy costs. In addition, if linked to job creation for the socially vulnerable, such as people with disabilities or low-income elderly groups, they can secure both justification for support as social enterprises and self-sustainability at the same time.”

To what level has research on energy-saving technologies progressed to overcome Korea’s ‘four-season climate’?
“Korea’s climate, where heat waves and severe cold alternate, creates a heavy burden of energy costs. The Project Group is conducting demonstrations with the goal of reducing energy use by more than 25% in real environments through system research using hybrid heat-source heat pumps and high-durability thermal storage materials. In addition, we are developing a ‘self-reliant model for artificial light supply systems’ that stores and uses electricity produced by high-output solar modules in an ESS (Energy Storage System). Through package technologies that combine greenhouse structures, crop growth, and AI control, we will fundamentally lower the burden of heating and cooling costs for farms.”

Has your past experience in space agriculture research at the University of Arizona and the Kennedy Space Center influenced your current work?
“During my time as a visiting professor in Arizona in 2004, I stayed briefly at the Kennedy Center and directly encountered space agriculture research based on closed ecological systems. It was such a specialized field that it did not shake my personal research philosophy itself, but I believe studies on resource recycling, sustainability, and plant responses in extreme environments different from Earth provide very useful ideas and feedback for the advancement of highly sophisticated future bio-industries and smart farm technologies.”

The multi-ministerial package project is set to end in 2027. How are preparations for follow-up projects progressing?
“Research outcomes produced with enormous budgets must not remain only as PDF reports after the project ends or become fragmented into individual technologies. Fortunately, planning for follow-up projects has now begun, centered on the Ministry of Agriculture, Food and Rural Affairs. We must secure the continuity of next-generation smart farm R&D by efficiently linking the empirical data and AI models accumulated on the integrated platform ‘KoSToP’ built by the Project Group to next-generation national and regional projects and industrialization R&D, while adding new issues that enhance global competitiveness.”

What suggestions do you have for narrowing the gap between technology and the field, and what kind of future agriculture do you envision?
“Farmers, academia, industry, and government must clearly define their respective roles and solve short- and long-term tasks harmoniously without overlap. Korea has a small territory and limited resources, so ultimately the only survival strategy is to move toward advanced technological agriculture. As AI is introduced in the future, the gap between advanced technology and farmers in the field will rapidly narrow from both technological and economic perspectives. Ultimately, I look forward to ‘pleasant future agriculture,’ in which humans and AI coexist and relieve the physical and mental stress of farmers, as well as sustainable rural communities.”