The use of artificial intelligence in agriculture

As the world population expands over time, agricultural modernization is the dominant approach of humanity to curb drought.

The various mechanical and chemical inventions of the 1950s and 1960s represented the Third Agrarian Revolution. Adoption of pesticides, fertilizers and high-yielding crop species, among other measures, has transformed agriculture and ensured a secure food supply for many millions over the decades.

At the same time, modern agriculture has emerged as the culprit of global warming, responsible for about a third of greenhouse gas emissions such as carbon dioxide and methane.

Meanwhile, inflation on food prices is reaching an all-time high, while malnutrition is rising exponentially. Today, an estimated 2 billion people suffer from food insecurity (getting safe, adequate and nutritious food here is not guaranteed). About 690 million people are malnourished.

The Third Agrarian Revolution may have driven its course. As we search for innovations to urgently launch the Fourth Agricultural Revolution, everyone’s focus is on artificial intelligence (AI). AI, which has evolved rapidly over the past two decades, has a wide range of technology that can handle human-like cognitive processes such as logic. It is trained to make these decisions based on information from vast data.

AI in agriculture

In assisting humans in fields and factories, AI can process, synthesize and analyze large amounts of data consistently and continuously. It can outperform humans in detecting and detecting disorders such as plant diseases and in estimating yields and climate.

In many agricultural practices, AI completely frees growers from labor, such as tillage (soil preparation), planting, fertilizing, monitoring, and harvesting.

Algorithms already control drip-irrigation grids, combining soil-monitoring robots with command fleets and weed-detection rovers, self-driving tractors and harvesters. The fascination with AI’s opportunities creates incentives to hand it over with more agency and autonomy.

He hailed the technology as a way to revolutionize agriculture. The World Economic Forum, an international non-profit organization promoting public-private partnerships, spearheaded AI and AI-based agricultural robots (known as “agbots”) during the Fourth Agricultural Revolution.

But in implementing AI faster and more widely, we can increase agricultural productivity at the risk of safety. In our recent paper, published in Nature Machine Intelligence, we considered the risks of implementing these advanced and autonomous technologies in agriculture.

From hackers to accidents

First, because these technologies are connected to the Internet, criminals may try to hack them. Interruption to certain types of firecrackers can cause huge losses. In the US alone, soil erosion costs US $ 44 billion (£ 33.6 billion) a year. This is increasing the demand for precision farming, including group robotics, which will help farms maintain and mitigate its effects. But these groups of soil-monitoring robots rely on interconnected computer networks and are therefore subject to cyber-destruction and shutdown.

Similarly, tampering with weed-detecting rovers releases weeds at a significant cost. We may also see interference with sprayers, autonomous drones or robotic harvesters, which disable any crop operations.

Beyond the farm gateway, with increasing digitization and automation, entire agrifood supply chains are vulnerable to malicious cyber-attacks. At least 40 malware and ransomware attacks targeting food manufacturers, processors and packages were recorded in the US in 2021. The US $ 11 million ransomware attack on the world’s largest meatpacker JBS is the most significant.

Then there are the accidental accidents. Before sending the rover into the field, it is instructed by its human operator to detect specific parameters and identify specific anomalies such as plant pests. It ignores all other aspects, either by its own mechanical limitations or by command.

This also applies to wireless sensor networks deployed in farms, designed to observe and work on specific parameters, for example, soil nitrogen content. Through careless design, these autonomous systems prioritize short-term crop productivity over long-term environmental integrity. To increase yields, they can apply excessive amounts of herbicides, pesticides and fertilizers to the fields, which can have detrimental effects on the soil and waterways.

Rovers and sensor networks may not work as well as machines do occasionally, sending commands to sprayers and agrochemical dispensers based on false data. And we are more likely to see human error in programming machines.

Safety over speed

Agriculture is a very important domain for us to implement powerful but inadequately monitored and often experimental technologies quickly. If we do, the result is that they intensify crops but weaken ecosystems. As we emphasize Our paperThe most effective method to treat accidents is assessment and prevention.

How we design AI for agricultural use and must have experts from different fields in the process. For example, applied ecologists can advise on the possible unintended environmental consequences of agricultural AI, such as depletion of soil nutrients or excessive use of nitrogen and phosphorus fertilizers.

Also, hardware and software prototypes should be carefully tested in monitored environments (known as “digital sandboxes”) before they can be more widely implemented. In these places, moral hackers, also known as white hackers, can search for vulnerabilities in security and security.

This precautionary measure will slow down the spread of AI a little. Also make sure that the machines that graduate the sandbox are sensitive, safe and secure enough. Half a billion farms, global food security and the Fourth Agrarian Revolution are in balance.

– Conversation

Read the original article here.

Leave a Comment