In the quest for sustainable agriculture and food security, AI-driven plant breeding has been heralded as a game-changing solution. The adoption of machine learning and generative AI has progressed at an incredible pace, transforming various industries - and agriculture is no exception. Genomic sequencing and machine learning are already being used by companies to identify crops with desirable traits, and climate resilient varieties of these crops, in pursuit of food security. This process accelerates the traditional plant breeding approach, which typically relies on time-consuming and resource-intensive phenotyping and crossing of plants.
Researchers at Norwich Research Park are looking to commercialize their AI-driven software, TraitSeq,for optimization of breeding processes. It uses RNA sequencing, machine learning and CRISPR technology to identify biomarkers, predicting the best performance of specific traits in crops in varying environmental conditions, and edit genes accordingly. This could improve cost-efficiency and the rate of development of crops.
Agrochemical giant Bayer has also developed new “precision breeding” technology that mimics traditional techniques using genomic sequencing and generative AI. These techniques can be used to suggest new varieties, designed to be widely adaptable and climate resilient; or tailor size, flavour profile and colour to consumer tastes.
Precision Breeding uses artificial intelligence (AI) technology to guide genetic changes and to access more data so scientists can quickly and accurately identify the precise changes needed to remove negative plant traits or emphasize positive ones. Ultimately, precision breeding results in the delivery of seed varieties tailored to growers’ unique field conditions years ahead of schedule.
Legacy of the Green Revolution
The Green Revolution, which took place in between 1950 and 1970, is often hailed as a pivotal technological turning point in the modernization of global agriculture. With its emphasis on breeding wide-adaptation, high-yield crop varieties, and usage of chemical fertilisers, and pesticides, it succeeded in significantly increasing global food production. However, it also had unintended consequences that continue to shape our agricultural landscape today.
One of the major issues arising from the Green Revolution was its environmental impact. The focus on breeding widely-adaptable crops, with little regard to the variable environmental conditions in which they would be grown, required the extensive use of chemical fertilisers and pesticides. Consequently, this led to soil degradation, water pollution, and biodiversity loss. The introduction of monocultures in place of indigenous crops also heightened susceptibility to diseases and pests, creating an unsustainable dependency on synthetic inputs.
In some cases, the Green Revolution also accentuated social inequality. While larger farms could easily absorb the cost of adopting new technology, reaping increased yields, smallholder farmers who invested in these new crop varieties often struggled to afford the necessary fertiliser and pesticide inputs, resulting in diminished crop yields compared to their previous crop varieties.
A vision of the future, or a hallucination?
AI-driven plant breeding could offer a unique opportunity to address the myriad challenges posed by conventional breeding methods in a changing climate. However, in an era where the rapid uptake of generative AI is raising urgent ethical questions about the use of this technology, it is important to diagnose the nature of the problem for which it is being presented as a solution.
One notable concern is that the current problem of food security is one of access, with global hunger prevailing despite rates of food production being more than enough to feed the global population. While technological advances are important, it seems misguided to present these advances as solutions to a problem which is primarily a result of economic shocks caused by conflict, with “70 percent of the world’s hungry people living in areas afflicted by war.” Climate change is a major driver of hunger, but climate adaptation of crops can not alone address the problems faced by populations unable to access food due to war and economic shocks.
There is also the pervasive issue of hallucination in generative AI, which presents notable problems for its integration into plant breeding. AI-generated outputs that deviate from reality are a concern particularly relevant to the field of biotechnology and plant breeding, where such hallucinations could potentially lead to, for example, suboptimal yields, inaccurate predictions of plant traits or unsuitable genetic combinations. This is all not to mention the concerns surrounding CRISPR and genetically modified food.
While the possibilities of AI-driven plant breeding have captured imaginations and received significant funding, it is essential for there to be democratic and participatory development of these technologies, as well as distribution of their benefits. Inattention to agro-environmental and socioeconomic conditions in central development of “wide adaptation” crops was a significant contributor to the negative impacts of the Green Revolution. As such, it is vital to ensure that all stakeholders, including smallholder farmers, researchers, policymakers, and local communities, have a voice in the development of AI-driven plant breeding, and ultimately how and if it is implemented.
As the impact of anthropogenic climate change threatens to undermine the global food system, it is crucial to mitigate the impact of this to safeguard food security and reduce global hunger. However, some proponents of using AI-driven plant breeding to address food security have fundamentally misdiagnosed the current problem as being primarily one of a shortage of food, rather than one of unequal access and distribution. For any such technological solutions to have a meaningful impact on food security in the future, they need to be developed in a democratized, participatory way that is accessible to smallholder farmers, who produce a third of the world’s food.
With respect to the question of climate resilience, AI undoubtedly holds great promise and potential for quick and cost-efficient adaptation to a rapidly changing global climate. However, if proposed technological solutions are not subject to intense scrutiny, there is a risk of exacerbating social inequality and environmental damage. Technological advances like generative AI and CRISPR have the potential to vastly improve human health and well-being, but these technologies are not neutral. Furthermore, if we do not yet trust in the capabilities of generative AI to provide us with accurate information without hallucination, we should at the very least exercise extreme caution when using it in the production and replication of biological material.