From farm to finance: How artificial intelligence can lead the change in rural India
There is no doubt today that artificial intelligence (AI) will, in the long run, be one of the most powerful technologies humans would have ever invented. The potential of AI extends beyond Wall Street and Silicon Valley, and now has become even more relevant for areas that have not been under the spotlight of development.
The largest use cases of AI have been in areas providing services to people out of reach of conventional technologies. The leaps that AI has taken in the fields of healthcare, finance, agriculture and law go on to prove that point.
For the longest time, healthcare has been a sector which is out of the reach of the less privileged. Access to doctors and hospitals in rural India is extremely limited. However, advances in technology have allowed the tide to turn in their favour.
To start with, using AI for the prognosis of diseases is already being done by tools like IBM’s Watson. Watson has tied up with Manipal University’s cancer department to assist doctors in providing individualised care plans based on research updated on a regular basis.
AI researchers at the Massachusetts Institute of Technology have developed computer vision algorithms which use machine learning (ML) to detect cancerous cells in biopsies at accuracy rates surpassing those of human radiologists. This is especially possible because of the large data sets available within the healthcare sector.
Further, the universality of medical conditions allows one to learn in one place and implement solutions globally. These emerging technologies are converging and will very soon allow us to have medical tests and initial prognosis possible, while sitting at home. The biggest benefit of this will be for the people living in rural India who might put off getting something checked due to the long travel required to get to the closest hospital.
Finance is another sector that is benefitting from the implementation of AI, primarily because of the existence of large data sets. Assessing risk is key to be able to lend. However, large parts of rural India have survived without access to any institutional financial services due to a lack of formal accounting practices and data. This, in turn, has led to a further dearth of data being created, which could have been utilised for providing optimum financial services.
However, advancements in data analysis has allowed companies to come up with unique risk assessment techniques. They assess financial risk not by looking at data related to finance, but by looking at more generic data like calls, SMSs, phone recharge patterns, browsing history, frequency and duration travelled, frequency of shopping. This data, which are plentiful in rural India, can be used to create a risk profile for a first-time banking customer.
The ML algorithms can learn patterns from data collected from people with access to banking services, to get a baseline on how these affect a true credit score. This algorithm can then be used to now provide credit scores to people without any financial data, who would previously be considered to be pariahs to financial institutions. This approach has allowed various small fin-tech companies to redefine the lending practices in India. These novel approaches, mixed with automated customer-relation services, like chatbots capable of recognising regional languages, will help financial institutions reach out to people in rural India, remotely.
Similarly, AI has also made improvements in the field of agriculture. Since agriculture is largely dependent on monsoon, technologies have already been created that allow or help farmers to predict weather patterns. This has been a relatively easy task since large quantities of data on the weather patterns, around an area, already exist. But the same cannot be said about soil and crop cultivation.
A number of Internet of Things companies are working to develop cheap solar-powered sensors that can detect and collect data on soil moisture and nutrient content. AI models, created using this data paired with historical monsoon and crop-output information, can predict expected yields based on weather, soil and other factors. Farmers can use these models to then decide which crops to grow, how much to water the fields and how to fulfil nutrient requirements.
Another agricultural application which is already seeing adoption in Europe is the use of a smartphone camera combined with computer vision to identify pests and diseases instantly, just by taking a picture. Early detection of pests and diseases is critical for farmers as it allows them to save their crops and maintain yield, ensuring a sustainable return from farming as a profession.
Lastly, and in a sector least expected, AI has been making substantial advancements in allowing people access to quality legal services. For a place like rural India, where most businesses are still conducted on handshakes, people face a huge risk of non-compliance as the enforcement of such agreements is close to impossible.
Further, the complexity of laws and the lack of consistency between different jurisdictions make implementation of technology in order to provide legal services even more difficult. However, if people could be given access to cheap and quality legal services, it could effectively help improve standards of living and economies of countries at large.
There are a number of startups and research organisations that are working towards making access to justice easier. Since, law is primarily a document-heavy profession, companies like tech firm Microsoft Corporation and Chinese search engine Baidu Inc. have demonstrated that AI programs can read and analyse documents and situations at human levels of speed and accuracy. There have been various similar advances over the last five years and startups are trying hard to democratise legal services, so that the public can access quality legal services that have, so far, been restricted exclusively to the elite.
For example, SpotDraft allows its users, who have no prior legal knowledge, to create, execute and operationalise contracts for free or at a fraction of the cost incurred for engaging with a lawyer.
Finally, rural India can also see a lot of job creation in the AI training space. Amazon’s Mechanical Turk and startups like Scale are providing human intelligence as a service. They get real people to perform repetitive tasks that computers are today unable to perform. The need for such workers to label, annotate and tag digital content is growing exponentially. These labelled data sets are required as input to train ML models.
People in villages with a mobile broadband connection can sign up and work on such tasks with minimal training. Even an uneducated human is an expert at tasks like object recognition and scene recognition, while even the most complex ML models can be fooled or confused by such tasks.
Having said that, there is little doubt that these applications of AI will improve lives. But, in parallel, the government should also ensure that rules and regulations are created in a manner that fosters innovation and does not deter it.
Job losses are a big debate topic, and a lot of hyperboles are used to outlaw things like self-driving cars. The policymakers should not look to delay the inevitable, but instead look for ways to reskill and prepare the people for the transition that will come regardless of bans taking place or not.
Madhav Bhagat is co-founder of legal-tech startup SpotDraft and ex-lead for Google Classrooms.