Agenda for AI Application in agriculture, health, environment & enterprise

AGRICULTURE

a. There is an enormous amount of data available on soil, plant, animal and human nutrition. However, we have not developed any analytical platform which can connect the variables for tracking eco-system health and ensuring preventive crop, animal and human nutrition.

b. The majority of the farmer machinery whether manual, bullock drawn or motorized do not have feedback control system. Most mechanical machineries have practically no electronic applications. Dynamic feedback from individual farm does not get aggregated at eco-system level and therefore knowledge based efficiency inputs are not available to farmers. Given the agrological diversity response variabilities of inputs such as water, seed and other chemical & non chemical inputs is not analyzed dynamically to generate mid-course correction.

c. Given large scale climatic fluctuation is super imposed on graphic and agro ecological variability requires development of dynamic contingency option. Such advice can be generated only with development of expert systems capable of using macro, miso and micro level agro meteorological data. Innovative coping strategies can neither be tracked nor disseminated without using AI applications, big data & deep learning.

d. The point of care devices for identifying pest and diseases for plants, animals and human require deep learning tools. While such devices for human use is being developed without always being contextualized. The probabilities of adverse consequences of different socio ecological and health interaction can be worked out better with the AI tools.

e. Point of care diagnostics and treatment advisory for animals has only recently began to develop.(ss.sristi.org) Since livestock distribution is much less in equities then land distribution application of AI tools will have much higher social welfare impact.

HEALTH

a. Leading public hospitals such as AIIMS & PGI get as many as 10,000 patients a day. No analytical tool is used to analyse the interaction between treatment, disease variability, gender, age and other variables. While intuitive understanding just grows among experts. In the absence of analytical tools, patterns cannot be identified easily.

b. In many small places, advance diagnostic facilities such as MRI and basic diagnostic tool such as X Ray are not available. India has reported less than 10,000 radiologists. A few startups have come up to convert 2D XRay into 3D images (Saket, IIT Bombay)and deep learning interpretation of XRay (IIT Madras).

The author is founder of Honey Bee Network & visiting faculty at IIM-A
anilgb@gmail.com