Although artificial intelligence (AI) has the potential to improve crop management and agricultural output, experts warn that there are significant risks associated with implementing new AI technologies that are being overlooked.
"The implications of machine learning (ML) models, expert systems, and autonomous machines for farms, farmers, and food security are little understood and underappreciated," according to the authors of a recent risk study published in Nature.
The researchers examined the risks of AI in agriculture, including interoperability, safety and security, data dependability, and unforeseen socio-ecological effects from the deployment of machine learning models to optimize yields.
What Role Does AI Play In Agriculture?
By quickly recognizing plant illnesses and administering agrochemicals efficiently, AI may be utilized in agriculture to improve crop management and yield. Machine learning can aid in quick plant phenotyping, agricultural monitoring, soil composition assessment, weather forecasting, and yield prediction.
However, according to Asaf Tzachor of the University of Cambridge's Centre for the Study of Existential Risk (CSER), the implementation of AI and ML design might jeopardize ecosystems and expose producers and agri-food providers to accidents and cyberattacks.
Disadvantages of Use of AI in Agriculture:
Before safely adopting AI for agriculture, the authors have highlighted a number of risks that must be considered.
According to the experts, cyber-attackers may contaminate databases and, among other things, shut off sprayers, autonomous drones, and robotic harvesters.
The reliability and usefulness of agricultural data is also a challenge, as indigenous farming methods are largely underrepresented in statistics, despite their significant contribution to local food security.
India's Concerns
Cognitive computing is being utilised in India to learn, comprehend, and interact with various contexts in order to increase productivity. Microsoft is partnering with 175 farmers in Andhra Pradesh to give agricultural, land, and fertiliser advisory services, which resulted in a 30% improvement in yield per hectare in 2016.
United Phosphorous (UPL), India's largest manufacturer of agrochemicals, has also partnered with Google to build a Pest Risk Prediction API that utilises AI to predict the risk of pest assault in advance.
In the initial phase, the app supplied automated voice calls for cotton harvests to roughly 3,000 marginal farmers in Telangana, Maharashtra, and Madhya Pradesh who had less than five acres of land. Based on weather conditions and sowing recommendations, the calls offered information on pest attack threats. One of the most serious concerns of AI in India is that it may expose such farmers to false information.
Furthermore, due to marginalisation, poor internet penetration, and a digital divide in India, smallholders may not be able to adopt such modern technology, expanding the gap between commercial and subsistence farmers.
What Can Be Done To Prevent This From Happening?
The researchers advocate enlisting the help of 'white hat hackers' in uncovering security holes in order to safeguard people from intrusions.
The dangers also highlight the need for "agricultural AI systems and services that are sensitive to context, taking into account potential social and ecological repercussions," according to the report.
Risks may be avoided if extensive risk assessments and governance mechanisms were implemented.
