Making agriculture smarter must be a COP28 priority

Author: Wendy Gonzalez. PHOTO/FILE/COURTESY

What you need to know:

  • We have to be ready to make agriculture smarter with the help of artificial intelligence (AI) and other technologies... 

About a third of global greenhouse gas emissions result from food production, distribution and consumption. That implies that any attempt to meet the terms set out in the Paris Agreement has to address this vital part of our global ecosystem if it is to have a chance of success.

Reducing greenhouse gas emissions and feeding a growing population are contradictory: more sustainable agricultural practices would, almost by necessity, reduce the amount of output in our food ecosystem because that ecosystem has been optimised for maximum output at all costs.

Agricultural technology in its myriad forms has to help bridge the gap and must be a key consideration in climate talks at the ongoing COP28 hosted in the Middle East and North Africa region, where food insecurity is quickly growing. 

We have to be ready to make agriculture smarter with the help of artificial intelligence (AI) and other technologies and perhaps more importantly, prepare to come together to support it across borders, governments and private companies.

Artificial Intelligence (AI) broad use cases in agriculture makes a clear case for more investment and collaboration in this area. Take pesticides as an example. Many of them, including sulfuryl fluoride (which is nearly 5,000 times as potent as carbon dioxide), are powerful greenhouse gases. 

In addition, 99 percent of synthetic chemicals, including pesticides, are derived from fossil fuels. 
This is before we get to biodiversity loss or other more visible effects of pesticide overuse. Studies have already found that we can reduce pesticide use by 50 percent without significant changes to crop yields — meaning there is a real opportunity to use pesticides more wisely and less wantonly than current practices. 

A computer vision model trained to differentiate crops from weeds or to identify different types of insects can enable more targeted application of pesticides, resulting in less greenhouse gases and environmental damage. It also reduces costs for farmers, thus improving their overall profitability. Finding more efficient ways to raise livestock is also critical. 

Current farming practices, especially when raising cattle, are some of the single biggest contributing factors to climate change.  Here, AI can help with precision feeding, which takes an individual animal’s needs into greater account. 

Granted, this technology can be expensive, and this is where government support could be most impactful.  

There will also be a plan for a credit guarantee scheme to provide financial support and training for start-up growth.  What if there were credits for investing in a precision feeding system or using an AI model to identify when to use pesticides? 

What if other subsidies were altered to encourage these practices or even regenerative agriculture? Of course, smart technologies like AI aren’t the only solution, and they have some caveats, with environmental concerns of their own — like energy usage.  We can’t zero out the gains made with AI in agriculture to pay for the added resources used in computing power to build effective models. 
Instead, we need to focus on processing only the data that has the biggest impact on model performance. 

In our work with startups and agricultural giants, we have seen the impact AI can have. What would happen if we all came together globally to help drive better understanding and, ultimately, policy changes around AI and agricultural technologies?

I think the answer is simple: We would make huge gains in the fight against climate change.  COP28 is the perfect time to look beyond governments, beyond NGOs, to all stakeholders, from farmers and major corporations to businesses like ours, in data validation and find a way for us to work together towards the same goal of a sustainably fed planet.

 Wendy Gonzalez is an AI and machine learning expert.