DIGITAL LIFE

In the field, AI helps detect crop pests and control livestock in Brazil

Artificial intelligence (AI) is already part of the routine of 41.9% of farms and agribusinesses in Brazil, according to an estimate prepared by Professor Oscar Burd, from the Getulio Vargas Foundation (FGV). In 2022, this rate was 16.9%. To arrive at this number, the specialist cross-referenced data from surveys by IBGE (Pintec Semiannual 2024), Sebrae, the World Intellectual Property Organization (WIPO), and consultancies specializing in agtechs.

— The speed is surprising. While previous technologies, such as GPS, took decades to become widespread, AI jumped from an experimental curiosity to a core business tool in less than five years — says Burd.

According to him, AI is no longer exclusive to large groups and has been "democratized" with applications, platforms, and technologies embedded in equipment:

— Today, low-power tractors already leave the factory with intelligent monitoring systems, and startups offer AI solutions as a service (SaaS), which allow small producers to access diagnostics via smartphone.

The technology has multiple applications in the daily routine of farms, such as in crop management, animal husbandry, and business management.

With digitalization, almost everything can generate data, which is processed by AI programs. From there, the technology relates them to available information databases and generates recommendations for productive activities.

Back to the 'good old days'... SLC Agrícola, one of the largest grain producers in the country, entered the "AI Era" between 2017 and 2018. According to the group's Technology Director, Rafael Rosa, AI permeates various activities, with applications and programs embedded in equipment (machinery, drones, sensors, scales, silos, etc.).

The main benefit, he says, lies in the analysis of large volumes of data to allow for faster decision-making:

— Just one tractor can generate a million data points per day. But, in addition to them, we have data from sensors, satellites, drones, and thousands of other pieces of equipment.

In agronomic management, the benefits include the detection of pests and diseases through real-time imaging for selective spraying, prediction of water stress per plot, and nutritional deficiencies in plants.

Rural producer Tasso Jayme sees in AI the possibility of returning to the “good old days of livestock farming.” He has 2,500 head of cattle in Goianésia (GO), about 170 km from Goiânia, in addition to 4,000 hectares of soybean and sugarcane crops.

— I've been in livestock farming for 50 years, since adolescence. Before, it was my main activity, but a few years ago it became complicated. New technologies can bring back the good old days — says Jayme.

He started using the technology in February: a drone with computer vision that identifies the best areas and the exact volume of seeds for planting pasture:

— With AI, we can locate the ideal areas to sow seeds over the soybean fields. This way, the pasture will benefit from the fertilization. The drone took images to indicate the needs for liming, fertilization, and the number of head (of cattle) per hectare — he says.

For the rancher, those who don't adopt AI will be left behind.

But there are structural obstacles. One of them is connectivity. Although 4G and 5G coverage in rural properties reached 43.8% in 2024, according to ConectarAgro, more than half of the properties still depend on offline solutions or satellite connections.

— Technology can be sophisticated, but without connectivity it simply doesn't reach the field — says Burd.

According to him, another obstacle is the scarcity of professionals capable of interpreting the data and recommendations produced by AI systems.

Unlike office AI, field AI focuses on physical autonomy and data collection in locations with difficult connectivity.

1. Areas of operation and impact...AI is transforming work "on the ground" through tools that increase productivity and safety:

-Precision agriculture: Use of sensors and AI to monitor crop health, predict harvests, and automate irrigation and planting.

-Predictive maintenance: Systems that analyze photos and sensor data to identify structural flaws (such as rust or cracks) in machines or infrastructure before they cause disruptions.

-Workplace safety: Smart cameras monitor construction sites to ensure the use of PPE and prevent accidents in real time.

Voice data collection: Tools like Fulcrum's Audio FastFill allow technicians to fill out complex reports simply by speaking, eliminating manual typing in the field.

2. Main technologies used...For an AI to operate outside the laboratory, it depends on three pillars:

-Edge AI: Data processing directly on the device (such as a drone or robot), allowing the machine to make decisions even without internet access.

-Computer vision: Allows machines to identify objects, obstacles, and hazards through cameras.

-Autonomous robotics: Software that allows vehicles to navigate irregular and unmapped terrain, an area led by companies like FieldAI.

3. The future...Experts predict that AI will cease to be just a tool and become a "digital colleague." In the field, this means robots that learn new tasks simply by observing humans and systems that manage entire fleets of machines in a coordinated way to maximize energy and operational efficiency.

mundophone