AI in the global south is often posited as a solution for ‘bridging the gap’ between developed and developing countries, and ‘leapfrogging’ stages in technological transitions. In the context of agriculture, it is expected that AI could vastly improve farmers’ lives, revolutionise agricultural productivity and enable sustainable agricultural practices. However, despite the promises of AI, the potential risks and unintended consequences of widespread AI adoption, and digitalization in the agriculture sector continue to remain under-explored. As the two speculative fiction narratives in this blog series shows, the implications of AI are likely to be contextual, unique, and regionally specific.
Understanding the social and human dimensions of AI is particularly important in the context of India, where the promise of AI and agritech are often touted as a way to uplift India’s ailing agri-economy, improve farmers incomes, enable access to credit and other forms market-driven advisory services. For instance, in its 2018 AI Strategy for India, NITI Aayog highlighted agriculture as a key sector for AI development. Similarly in 2022, the Prime Minister announced the integration of a Data for Development (D4D) principle into India’s policy planning. Further, In 2020-21 and 2021-22, the Indian Government allocated INR 1756.3 crores and INR 2422.7 crores for state governments to introduce new technologies including drones, artificial intelligence, block chain, remote sensing and GIS etc in agriculture. In March 2023, the Ministry of Agriculture & Farmers Welfare also announced the “Kisan Drones” initiative for up to 10 lakhs in subsidies to buy drones.
Within the private sector as well, AI and agritech start-ups have rapidly grown with more than 1,000 agri-tech startups in India that use AI to optimise supply chains, provide agri-advisory, predict demand, and enhance market efficiency. Through the WEF’s Artificial Intelligence for Agriculture Innovation (AI4AI) initiative, the Telangana government was able to launch the Saagu Baagu pilot, to use AI for pest and disease detection, soil analysis, and precision farming techniques in chilli and groundnut cultivation. Digital Green and ColoredCow released a chatbot that uses AI to provide farmers with customised notifications and short videos on a real-time basis, helping them plan and manage their crops more efficiently. Agnext developed an AI-based food quality assessment technology to simplify the process of on-spot quality standardisation as well as fostering economic, social and ecological profits. In Andhra Pradesh, under the Rythu Bharosa initiative, farmers will have access to InSight, an AI-based quality assurance app for farm inputs. Tamil Nadu’s agricultural budget in 2022, mentioned that AI would be utilised to monitor the pest and disease infestation in crops and SMS advisories would be sent to farmers on instantaneous management measures. Microsoft India has signed an MoU to leverage Microsoft’s cloud, AI, research, and industry knowledge to create new solutions, particularly in agriculture.
Thus, a combination of policy interest, investments and pilot experiments in AI and growing ecosystem of AI tech-companies and start-ups in agriculture, is providing a heady impetus for change.
However, what these changes might mean for the millions of small and marginal farmers, farmhands and landless labourers employed in Indian agriculture is far from clear, and could end up constituting the fine print in-between the grand promises of AI. While on the face of it, the implementation of AI could lead to some productivity gains, and improvements in farmer incomes, the social dimensions of a transition towards extensive digitalisation and ‘AI-fication’ of the Indian agriculture landscape remains grossly under-explored. Questions arise regarding the accessibility and affordability of automation technologies for small-scale farmers, the potential displacement of labour in rural communities, and the concentration of power and data in the hands of a few dominant actors.
A variety of metrics show that India’s agriculture sector is ailing. Farm debt per household has increased by 57.7% in 2018 since 2013 according to the National Statistical Office. Nearly one in four farmers in India live below the poverty line. As farmers struggle under income and regional disparities, climate change also threatens these meagre earnings. India has seen significant increases in temperature, frequent heat waves, droughts, extreme precipitation events, and intense cyclonic activities. This is coupled with a reduction in rainfall, a primary water source for agriculture. Research suggests that a lack of information, followed by household income, farm size, and credit accessibility, are the major influential factors in adopting the requiered adaptation measures.
While there have been significant gains in terms of data and research on AI use in India and in the global south, such research pales in comparison to the studies that have been conducted in the global north. In an Oxford study on AI preparedness, 193 were ranked and India was ranked 40; the lowest-scoring regions included much of the Global South. In the global north, research highlights the potential for bias, discrimination, algorithmic determination etc. in the application of AI systems, however, regional, social and cultural differences continue to be critical blindspots in understanding the implications of AI. For any solution to work, it has to be contextually appropriate so global insights cannot be applied loosely here.
In the light of the technology-driven transitions that seem to be underway in Indian agriculture, we have developed two specualtuve fiction case-studies of AI, highlighting the potential pitfalls and challenges in widespread adoption of AI and other emerging digital technologies. These fictional narratives are not meant to predict the future, rather, they aim to serve as crucibles for critical thought, empathy and imagination. They highlight the need for a nuanced understanding of the assumptions and values embedded within AI systems and close attention to the power dynamics at play, ensuring that AI does not become a tool for further marginalisation and dispossession. We believe that technological transitions alone will not address societal concerns, as these technologies do not autonomously tackle the structural dynamics of power imbalances and social inequities prevalent in society. Further, the emergence and development of technology is a social process and technological trajectories need to be socially shaped, their values aligned to societal needs, and well-being in order to achieve just and equitable futures.
The need of the hour, therefore, is to contextualise AI technologies and their diffusion, use and governance. Future foresight methods like that of speculative fiction can be used to provide insights and ways of thinking/knowing that can help tailor effective policy solutions. Given the dearth of research on the outcomes of such AI-based tech deployments in the global south, India needs a new approach to AI adoption in agriculture. A confluence of diverse research methods centering climate resilience, social empowerment, and public-private innovation policies, along with technology design strategies that respond to the needs, aspirations, and behavioural practices of farmers are needed.