This section discusses potential future trajectories of agricultural automation technologies for different types of countries and farms, in the light of structural factors that can shape the dissemination and adoption of these technologies. It looks at prospects for making mechanized agriculture more sustainable. The benefits of motorized mechanization have brought some negative environmental impacts, inter alia, crop land expansion taking place at the expense of forests or pastures of savannah land.50 Furthermore, it discusses the potential for automation of small-scale agricultural production and some of the economic and social implications of future automation trajectories.
Prospects for highly mechanized agriculture to become more sustainable
In high-income countries, but also on many commercial farms in low- and middle-income countries, agriculture is already highly mechanized, partly in response to scarcity or seasonality of agricultural labour. For economies of scale, large machinery is mostly used. However, evidence shows how this has caused soil erosion, deforestation, increased greenhouse gas (GHG) emissions and biodiversity loss.51 In many countries, service providers often use large machinery and mostly serve those farmers who have cleared trees and tree stumps from their plots;40, 52 but the removal of farm trees and altered cropping patterns triggered by mechanization can contribute to soil erosion.7 Moreover, the soil erosion and degradation caused by heavy large machinery also leads to declining yields.38, 53 Using large tractors has fundamentally changed the face of rural landscapes as producers often enlarge and reshape plots, leading to a loss of farmland diversity and biodiversity for food and agriculture.50, 52 Motorized mechanization is associated with reduced crop diversity as it encourages production shifts towards easier-to-mechanize crops, such as wheat, maize and rice.4 Regrettably, farmers often do not adopt biodiversity-enhancing practices, such as conservation agriculture, intercropping and rotations, because they are very labour-intensive.54 Mechanization often leads to more specialization and less commodity diversification, and this can reduce resilience.55
To address these challenges, innovations in motorized mechanization can be tailored to smaller, lighter machinery which can reduce soil compaction and mitigate negative environmental impacts. Scale-appropriate automation adapted to local conditions can play an important role in reducing those effects. Autonomous robots can help reduce chemical and energy use, as well as GHG emissions if powered by renewable energy.56 Applied technical and agronomic research can help explore mechanization solutions that best fit local agroecological conditions. Governments can also use policies to promote access to machinery and equipment proven to be more environmentally friendly.38, 40
Conservation agriculture can reduce soil erosion, using rippers or direct planters to replace ploughs. Coupled with crop rotation and permanent soil covers, these minimum soil disturbance practices can reduce soil erosion by up to 99 percent.57 Conservation agriculture appears to be a way forward for agriculture, but locally adapted solutions are needed to avoid some of the challenges.58 In this context, in May 2019, regional training on appropriate mechanization for conservation agriculture was co-organized by the Centre for Sustainable Agricultural Mechanization – a regional institution of the United Nations Economic and Social Commission for Asia and the Pacific – and partners in Cambodia.59
Transitioning to renewable energy is important from not only an environmental perspective but also a financial one. The case studies of TROTRO Tractor in Ghana and Tun Yat in Myanmar identify increasing and unstable fuel prices as important barriers to adoption (see Annex 1). Moreover, renewable energy offers new potential to power automation along the value chain and may be particularly attractive for remote rural areas.60 However, not all operations are efficiently run using currently available renewable energy sources. For example, electricity is not suitable for power-intensive land preparation. Research is needed to explore which off-grid renewable energy solutions can most efficiently power each type of machinery along the value chain.51
Chapter 2 showed that labour shortages, as well as the need for increased efficiency and resilience to climate shocks and stresses, are driving adoption of digital automation technologies and robotics with AI on highly mechanized farms. The evidence points to environmental benefits from these technologies and can be useful for guiding future innovations; however, given the limited data and the fact that many solutions are still in the early stages of development and commercialization (see Figure 6), it is not possible to generalize about the potential benefits. As these technologies are further developed and more widely adopted worldwide, including through shared use or hiring services, adoption may expand to smaller-scale farmers.31
In high-income countries, robots are replacing manual labour in tasks ranging from irrigation, pest scouting, harvesting and weeding, to fruit selection and picking. For example, in one case study (Harvest CROO Robotics), the service provider noted that 70 percent of strawberry producers in the United States of America have already invested in their project to develop strawberry harvesting robots (see Box 16). Robotics technologies can lead to environmental benefits if they reduce or eliminate the use of pesticides and herbicides. Autonomous crop robots save labour, improve the timing of operations, optimize quantities of applied inputs, and reduce soil compaction, especially when using smaller swarm robots. Based on a review of 18 studies, autonomous crop robots for harvesting, seeding and weeding are economically feasible in certain circumstances.61, 62, 63 Swarm robots, in particular, offer a cost advantage on farms with small, irregularly shaped fields.64 Policymakers and producers need to gain a clearer perception of these benefits in order to achieve increased investments in development of the relevant technologies.
BOX 16Solving labour shortages in strawberry fields using harvesting robots
Automated harvesters can autonomously pick, inspect, clean and pack crops. Harvest CROO Robotics was developed in the United States of America to solve the problem of labour shortages in the strawberry production industry through a robotic harvester. Each harvester has 16 independently working robots, which navigate through the farm, inspect the quality and ripeness of the strawberries, and then proceed to pick, clean and pack them. This technology thus completely replaces manual labour in the diagnosis, decision-making and performing of this task.
Harvest CROO Robotics is one of the few known strawberry harvesting solutions currently available in the United States of America. It has attracted investment funds from about 70 percent of national strawberry growers – typically large-scale – in response to concerns related to both lack and cost of labour. A pay-as-you-go system is adopted, with producers’ payments related to the volume harvested.
Once the technology scales, the aim is to have a fleet of harvesters that can be controlled remotely from an operations centre; in addition to picking, inspecting, cleaning and packing, it will also be possible to collect data to be shared with the growers.
SOURCE: Ceccarelli et al., 2022.31
The potential of automation for unmechanized or scarcely mechanized small-scale agriculture
Small-scale agricultural producers comprise a highly diverse range of agricultural production units. Some may be highly commercialized and use modern technologies, including motorized mechanization, while others practise subsistence farming with simple tools. In general, they rely heavily on family labour and mechanize only part of their farm operations – if at all. In many contexts, however, they could benefit from the expansion of rental machinery markets. The rental market tends to be dominated by large machinery that migrates across various agroecological zones within and across national borders. In order to take advantage of these services, producers have had to adapt their farms and production systems to conform to this focus on large-scale agricultural production. There is an urgent need, therefore, to find tailored solutions to first, address past negative impacts of mechanization and second, facilitate its expansion, thus increasing productivity in a sustainable manner.
Small machinery is a better fit for small-scale agriculture
Technological solutions such as small two-wheel and four-wheel tractors were key for increasing mechanization in Asia.2, 19, 20 Two-wheel tractors are likely to be more profitable and better adapted to small farm sizes. They can manoeuvre around tree stumps and stones and minimize biodiversity loss since they do not require substantial field clearing. They are also easier to operate, maintain and repair, and are more suitable for microfinance.22, 65 The same reasoning can be applied to a wider range of small agricultural motorized machines, which are more biodiversity-friendly as they do not require substantial reshaping or clearing of agricultural fields. There may also be benefits in terms of gender equality (see Box 17, which presents successful examples of small motorized machines used by women in Nepal), with potential savings in labour and resources, and an increase in women’s empowerment.
BOX 17The business case for women adopting motorized mechanization: evidence from Nepal
There are three ways in which motorized mechanization can empower women and respond to their needs. Women may be: (i) customers of mechanization service providers – reducing the drudgery of farm work and freeing up time for resting or other social or economic activities; (ii) operators of machinery and equipment or staff in a mechanization hiring business – using their technical skills to earn an income; and (iii) entrepreneurs managing their own mechanization hiring services agribusiness – providing mechanization services to other farmers and generating revenue.
A report recently produced by FAO provides information on market-tested machinery and equipment for crop production and post-harvest operations in Nepal. The goal is to promote and support women’s access to motorized agricultural mechanization as operators and/or managers. Examples of motorized equipment adopted by women include the following:
- The power weeder comes in several types and sizes. It performs weeding and interrow cultivation of wide-spaced crops such as vegetables, maize and sugar cane. According to the report, compared with manual labour, a single machine can weed a very large area. Women maize farmers in Dang district reported they could save NPR 10 000 (USD 84) per bigha (an area corresponding to 0.66 ha) by using the large power weeder rather than paying for manual weeding.
- The mobile thresher is an engine-powered thresher used for bundled rice or wheat. It eliminates the drudgery of threshing by hand, saves time and greatly increases the amount of grain threshed (8–10 times more than manual threshing). Due to its high threshing rate, it is suitable for individual service providers or custom hiring centres.
- The maize sheller is used to separate the grain from the cob. It eliminates the drudgery and pain of shelling by hand, saves time and greatly increases the amount of grain that can be shelled in a given time (30–40 times faster than manual shelling). Shelled maize grain also occupies less space than maize on cobs, making its storage easier.
SOURCE: Justice, Flores Rojas and Basnyat, 2022.66
Digital automation technologies can provide multiple benefits, but there are many challenges for small-scale agriculture
The increasing research on precision agriculture in low- and lower-middle-income countries highlights the need to harness the potential of digital automation technologies for small-scale agriculture.67, 68, 69 To spur adoption, some service providers consider offering free advisory services to small-scale producers, basing their business model on the potential income generated by selling the data collected from farmers.31 This option may be an encouraging starting point, provided it meets data sharing and privacy standards. Furthermore, there is willingness among farmers to grow the same crop on adjacent areas, to share the payment for UAS-based advisory services (e.g. in Burkina Faso,70 Ghana71 and Rwanda72).
Digital technologies have also provided a boost to agricultural advisory services for small-scale producers.30 In low-income countries, the most frequently deployed digital solutions are disembodied digital tools, due to their low cost, but impacts on productivity and environmental sustainability are still largely unknown. Moreover, the available data are insufficient to generate the tailored advice needed by small-scale producers. Further, the low level of digital skills leads to difficulties related to scaling, and there is also a strong digital divide with women and other vulnerable groups having poorer access to solutions. Another emerging issue in many countries is the absence of data privacy and protection legislation, which may lead to misuse of data by third parties.73
There is also research on using drones to apply inputs such as fertilizers and chemicals on small farms (including in Africa);74, 75 commercialization has begun, but the applied solutions are mostly map-based with very little autonomous decision-making capacity. Benefits of drone-based input application include improved precision, reduced pesticide exposure, the possibility of application in fields unreachable by equipment (because the field is too wet or difficult to access), and avoidance of damage to standing crops due to moving equipment. Profitability depends on the cost of equipment, effectiveness of application, savings in inputs due to spot application, and improved yields through reduced damage compared with use of ground-based machines. The availability and affordability of drones are key for small-scale agricultural producers, who do not usually own their own equipment. These technologies entail many challenges, such as refilling the spray tanks, fertilizer bins or seed hoppers, recharging batteries, employing pesticide labels for spot application, training users, and managing drift to non-target areas. Overcoming these issues requires technical and institutional capacities, and this in itself can be an additional challenge in many low- and middle-income countries.76
Given that one of the barriers to adoption of digital automation by small-scale producers is cost, improved technologies, scale and innovative business models are especially important to enhance affordability. This issue is clearly illustrated by computers and smartphones: once manufactured in large volumes, they became much less expensive, paving the way for their increased use in precision agriculture.31 In some contexts, water scarcity is a challenge to agricultural production; in Mali, a case of successful adoption of automated greenhouses (where a computer controlled water and pesticide applications) shows that such technologies can lead to greater efficiency in water and input use.77
Precision livestock farming
Precision livestock farming is mainly applied in intensive systems in high-income countries where sensors monitor the health, reproductive status and behaviour of animals. Electronic tagging and blockchain are increasingly used to improve product quality by facilitating traceability of livestock marketed from extensive systems.29 However, these advanced technologies are still too costly for most livestock producers in low-income countries, where precision livestock technologies focus more on virtual fencing systems with audio alerts, electric shocks or other prompts to keep animals within boundaries. These technologies reduce drudgery and labour requirements, facilitate reproductive management, collection of information and intensive management, and potentially eliminate the need for physical fencing. In addition, GNSS help livestock producers locate animals grazing in large open pastures; they can be linked to sensors to monitor temperature, movement and other indicators of health and reproductive status. Yet, individual GNSS for each animal are currently too costly for extensive grazing systems. As with crops, both re-engineering (to achieve lower costs and mass production) and innovative business models are needed to make these technologies available for extensive livestock production systems in low-income countries.29 Apps for accessing useful information related to livestock management offer great potential for precision livestock farming.78 There is anecdotal evidence from Kenya that pastoralists are increasingly using such apps to indicate the state of grasslands and help find sufficient feed when moving around with their herds.79 Satellite data-based apps can help determine and report animal diseases, allowing livestock producers and raisers to make rapid, targeted interventions.78
Asset-sharing arrangements for mechanization
Digital tools also hold great promise for improving asset-sharing of agricultural mechanization for small-scale producers. For example, GNSS tracking devices and fleet management software – such as those following the Uber-type solutions for ride-hailing – promise to significantly reduce the transaction costs of small-scale producers and machinery service providers and can facilitate the supervision of machinery operators by service providers.29 Examples include TROTRO Tractor in Africa, and Tun Yat in Asia. These initiatives face various challenges, such as poor roads and connectivity, and the fact that demand is seasonal, peaking in specific periods. Service providers are considering the use of institutional innovations to overcome some of the challenges. For example, by using booking agents to pool small-scale farmers, they reduce the transaction costs of reaching out to farmers and ensuring business continuity.80 Potentially this can allow progressive adoption of GNSS for accurate positioning and advanced machine control, with the additional prospect of further developing precision agriculture through VRT also in low- and middle-income countries. The main challenge to the use of GNSS with large machinery is the need for fields to be rectangular, which may not be the case for many small-scale producers.
Robots with artificial intelligence
Robots designed for farms in high-income countries are often not suitable for low- and middle-income countries, where farming is still dominated by small-scale producers relying mostly on family labour and performing many operations manually. For example, automated cotton harvesting machines in high-income countries are highly efficient but only suited to cotton that matures all together at the same time. This is because the machine can damage the plants while harvesting. Such a solution does not fit traditional farms in India or Western Africa where cotton is a high-quality, multi-bloom crop with a season lasting about 150–160 days, during which cotton is picked three to four times.31
Costs are an additional barrier to adoption, especially for small-scale producers in low- and middle-income countries, where very few examples of robotics solutions are found. These target crops and cropping systems are traditionally designed for manual work, and are tailored to local contexts and challenges, requiring minimal to no change in the current farm structures. The drivers of adoption of these solutions are also socioeconomic, with lack of seasonal labour being a prominent one. Other factors leading to the diminishing interest in manual, poorly paid labour include better access to education, migration to cities, social stigma and government policies to support the jobless.73, 81, 82, 83
The literature suggests that autonomous robots designed specifically for conditions in low- and middle-income countries bring the following potential benefits: (i) reduced human labour requirements; (ii) lower costs and reduced economies of scale, ensuring technologies are also accessible to smaller farms using conventional mechanization; and (iii) ability to use technologies in irregularly shaped fields in a cost-effective manner, thus avoiding the need to reshape rural landscapes into large rectangular fields on which traditional mechanization is most efficient. Unfortunately, there are no feasibility analyses for these countries to support the business case for investing in these technologies.29 This is in part due to the fact that the organizations developing these solutions lack the capacity to attract or retain talented personnel who can conduct such an analysis; they are generally small enterprises competing against large companies.31 Box 18 presents potential opportunities and challenges when developing robots for small-scale producers.
BOX 18A vision for low-cost autonomous crop robots
An example of a potentially feasible robot for small-scale producers would be a small-wheeled autonomous crop robot that can seed, weed and harvest, and costs the same as a motorbike (USD 500–1 000) – something that many agricultural households in low-income countries own and therefore serves as a useful price point. A leg robot could also be useful in fields, since it can step over obstacles, but it is much more expensive. Given the autonomous crop robot’s ability to learn using artificial intelligence (AI), there is great potential for a substantial increase in food production, far beyond currently feasible levels. However, producing specialized robots for each crop and for specific agroecological conditions is a high-cost, low-volume business. Therefore, a plausible business model is that a manufacturer delivers a generic autonomous machine, with a range of tools adapted to different tasks, some of which locally manufactured. The autonomous machine would be fitted with a global navigation satellite system (GNSS) device to allow it to create maps (e.g. soil colour, soil strength based on force required for hoeing, yield). It could be powered using various energy sources (e.g. fuel, solar, methane). To make such autonomous machines more affordable, especially in the early stages, they could be hired out or fees for service farm work could be charged.
With a generic autonomous crop machine, many other types of digital automation become possible. For example, with the incorporation of a crop sensor, the autonomous machine might determine fertilizer needs,84 make use of use previously recorded soil, plant and yield maps, and identify pests, diseases and weeds, applying insecticides, fungicides or herbicides as needed.
While it will be challenging for small-scale producers to access digital automation, at the same time millions of them represent a business opportunity and an enticing, new market. A similar process of research, technology development and entrepreneurship occurred with hermetic grain storage throughout Africa and Southern Asia.85 Prior to the Purdue improved Crop Storage (PICS) bag, manufacturers were reluctant to invest in grain storage innovations for small-scale producers because of their perceived lack of buying power. However, once PICS had sold millions of bags in more than 30 countries, many imitators and competitors emerged.
SOURCE: Lowenberg-DeBoer, 2022.29
Broader implications of digital automation technologies for agriculture
Agricultural technologies often have economic, social and environmental implications that extend far beyond their farm-level benefits and costs. For example, motorized mechanization of agriculture has often been associated with increased farm sizes, reshaping of fields and declining rural populations. Digital automation technologies have great potential to address the environmental challenges in highly mechanized agriculture as discussed above. If well-tailored, they also have great potential in small-scale farming, especially if combined with adapted motorized machinery. Looking further into the future, if the digital automation technologies discussed in this chapter, including robots and AI, are well developed and widely adopted, they could lead to broader positive implications including the following:
- Farm structure: Small swarm robots allow a reduction in economies of scale and eliminate incentives to expand farm size, thus avoiding social and environmental disruption. By reducing drudgery, increasing profitability and enhancing the reputation of agriculture as a high-tech industry, swarm robots can help rural communities retain the young and also attract workers from other sectors (see more on youth in Chapter 4). Motorized mechanization agriculture has led to the abandonment of small irregularly shaped fields; swarm robots could allow commercial agriculture to reclaim some of these abandoned fields, which are often characterized by good soil quality, reliable rainfall and proximity to markets. In turn, as swarm robots help improve the profitability of such fields, subsidy programmes directed at small-scale farms may become less costly. Furthermore, both small-scale farms and larger farms still relying on animal traction may be able to leapfrog motorized mechanization and adopt directly digital automation, avoiding the need to reshape rural landscapes, and thus contributing to greater biodiversity.
- Agricultural equipment market structure: Ensuring access for small and medium-sized agriculture – including crops, livestock and aquaculture – to various digital automation technologies may lead to changes in the structure of the associated equipment market. This can create opportunities for entrepreneurs who have the technical capacity to develop affordable, reliable, autonomous machines and equipment, and link that technology with innovative business models.
- Crop protection as a service business: Crop protection currently focuses mostly on selling large quantities of pesticides. Targeted spraying may reduce the quantity used by as much as 90 percent, with significant environmental benefits, while mechanical or laser weed control could eliminate herbicides entirely.29 This could strengthen the role of local entrepreneurs who provide standardized autonomous machines to identify weeds and pests. These machines might be provided under a fee-for-service model or sold directly to farmers.
- Safer and more efficient and resilient livestock and aquaculture: Digital automation can significantly facilitate remote work and help minimize the work burden, while improving also management.86 There is increasing research on the potential uses of digital technologies in aquaculture and on how the sector can make significant shifts in associated business models and farm structure.87 For example, IoT technologies can automatically monitor water conditions and allow fish farmers to take immediate action.88 In livestock production, the growing use of biometric sensors, which monitor an individual animal’s health and behaviour in real time, allows producers to obtain real-time information and thus perform targeted actions that can deliver many benefits, including reduced use of antibiotics. Sensors also enable blockchain technology, which can guarantee traceability of animal products from farm to table and provide key advantages in monitoring disease outbreaks and preventing associated economic losses and food-related health pandemics.89
Other implications will emerge as the technologies evolve and become more accessible. The exact implications will, however, depend on many factors, including technology characteristics, connectivity, legal and regulatory frameworks, business decisions by corporations and start-up companies, social media reactions, and cultural attitudes to agricultural digital automation. Governments can promote adoption and enable positive outcomes through digital infrastructure, appropriate legal and regulatory approaches, research, and education (see Chapter 5).
