Published :
Updated :
For all latest news, follow The Financial Express Google News channel.
We wait for summer months to enjoy mouth-watering delicious fruits, including watermelon. But it's a tricky job to spot optimally ripe watermelons. A mobile app with artificial intelligence could help in picking the right one, avoiding wastage. On an average, globally we waste one-third of the food produced for human consumption. Every year, we roughly waste 1 trillion dollar worth of food amounting to about $680 billion in industrialised countries and $310 billion in developing countries. Already, produced food is good enough to feed 10 billion people, although 1 billion out of 7 billion fellow human beings remain hungry due to lack of food. This is because, we waste. Solving food waste not only means meeting the nutritional needs of the world's population but it also addresses significant issues of carbon emissions. According to a United Nations' report, food thrown away represents 3.3 billion metric tons of carbon dioxide. It has been found that technology along the whole value chain from farming to placing food on the plate has huge potential in reducing the food wastage. Artificial intelligence technologies like computer vision, machine learning and data analysis have the key to improve the chain from the farm to trucks to stores in an effort to reduce food waste.
Among the 17 goals of sustainable development, goal number 2 focuses on food security, ending hunger, improving nutrition and promoting sustainable agriculture. Although emphasis has been on increasing the yield to feed the growing population, reduction of wastage of already produced food could be a useful means to meet the goal of providing quality food to the growing population. Moreover, reduction of food wastage will also increase the sustainability of agriculture and reduce greenhouse gas emission.
There have been diverse factors behind food wastage. One of them is weak forecasting of food production and demand. Often farmers produce far more food of certain types in a season outstripping the demand, resulting in wastage. During the harvesting of perishable foods such as vegetables and fruits, due to the weak short-term forecasting of demand and harvesting rate of fellow farmers, often farmers end up over-harvesting in a particular day. Real-time sharing of this information would encourage farmers to optimise their decision to more uniformly distribute their harvesting activities, resulting in less wastage and also higher income. To reduce food wastage by supermarkets due to poor forecasting, a start-up in the UK has developed AI application with the claim of making 50 per cent more accurate than traditional forecasting methods and can boost gross margin up to 18 per cent-by helping managers on planning optimal orders for perishable food. On the other hand, some supermarkets are using computer vision to automatically inspect the products to assess freshness and also price them accordingly to make sure that the food is sold before the expiry of shelf life. Some start-ups are working on computer vision-guided food cutters to reduce food wastage-as much as 10 per cent.
With increasing urbanisation, the distance between food production and consumption has been constantly increasing. The challenge of deciding about the right amount of harvest and of delivering harvested fresh food to consumer's plate has been constantly growing. The delivery of perishable food relies on what's known as the food cold chain. This vastly complex distribution of food from farm to plate relies on maximising the quality and longevity of crops. Adding artificial intelligence technologies like computer vision, machine learning, and data analytics for forecasting demand, assessing harvesting pattern, performing food inspections, trucking and refrigeration is bringing efficiency and effectiveness to food logistics. But many of those potentials are untapped, particularly in developing countries. It offers the opportunity to new generation AI start-ups to nurture diverse ideas into profitable innovations so that farmers, logistics service providers, traders, and food stores, among others, can profitably succeed in reducing wastage in delivering fresh and safe food to consumers, preferably at a lower price than before.
Nurturing AI start-ups in offering innovations for reducing food wastage through creating increased profitability of all stakeholders is a major challenge to benefit from this untapped vital potential-reducing hunger through reduction of food wastage. Despite the high potential of AI innovation in reducing wastage, there are indeed challenges of running into loss-making. It would be a long journey in improving underlying technologies, adding complementary ideas, creating the customer base to benefit from the scale and scope, and also stimulating the externality effect to reach profitable delivery of innovations. In addition to entrepreneurial aspiration, there should be a strong research and development support to keep improving innovations resulting in graduation to a profitable state. Often such R&D capacity would be beyond the reach of the individual entrepreneur. Moreover, financing such a long journey would be a critical issue. It's likely that the many high potential AI ideas for reducing food wastage would require 5 to 10 year long loss-making journey. It's likely that at the beginning, innovations would not create strong appeal among a critical number of customers in generating profitable revenue.
Inadequate research and development capacity and limited risk capital financing appear to be major barriers to exploit AI innovation potentials to reduce food wastage. In establishing relations between unknowns, AI innovations demand expertise far more than knowing programming languages and basics of computing. One of the major R&D challenges in leveraging this opportunity is data - high quality, reliable data. Access to vast quantities of data is vital for AI to be effective for reducing food wastage. Data annotation is crucial, given the requirement of tagged or annotated data as a prerequisite for machine learning in AI innovations. In most cases, data needs to be annotated manually or in semi-automated ways for the purpose of machine learning, even though sometimes annotated data can be generated automatically from the source. Measures need to be taken in annotated data production and sharing, creating a common market of data for supporting AI innovations. To address this vital issue, partnerships should be developed among policymakers, consumers, start-ups, large companies, farmers, department of agriculture, agricultural input providers, traders, logistic service providers and universities-among others.
The staggering amount of food wasted is a major concern in reaching the target of a hunger-free world. AI innovations can play a vital role in reducing this wastage. Starting from long-term to near real-time demand forecasting to machine vision based food inspection and optimum cutting, AI innovations can reduce food wastage along the whole value chain starting from production to harvesting to retailing. To capitalise on this, R&D ecosystem along with annotated common data market should be created to facilitate start-ups.
M Rokonuzzaman Ph.D is academic and researcher on technology, innovation and policy.