The world is undergoing a revolution in Artificial Intelligence (AI). If you’ve ever wondered how Gmail knows which ones of your emails are spam, AI systems are using machine learning algorithms to analyze large sets of existing data to filter your incoming emails. Machine learning is a subset of AI in which machines learn to complete tasks without being explicitly programmed. Without us noticing, AI systems are permeating many of the industries and services we use today. AI enables e-commerce platforms to make product recommendations, firms to provide legal advice online, and even police to detect the likelihood of crime in certain areas.
All this is possible due to systems that are trained to find patterns in large datasets and make predictions. The shift towards a world where AI plays a crucial role in its functioning is only a matter of time now. This means a social and economic change more pronounced than the one that took place after the industrial revolution. As a low-income country on the lower rungs of the industrial ladder, Nepal will face unique challenges as AI shapes societies and economies globally. The purpose of this article is to contextualize the opportunities and challenges of AI and offer possible strategies for future.
Economy and Employment
The ability of AI to use a large amount of data for analysis is enabling efficient use of resources in companies via automated planning, workflow optimization, supply chain and logistical optimizations, and the like. In Africa, micro-lending companies are using AI to define risk measurements for people who don’t have a credit history. The same companies are using AI to also detect fraud. This would be applicable to lending agencies in Nepal as well.
Moreover, the gaps in cognitive and socio-emotional skills when it comes to digital literacy means Nepali citizens would benefit from the natural language processing capabilities of AI systems to provide services like mobile banking, text-based customer support, etc. Small business owners can leverage AI automation to optimize business processes or create new business models. For the government, AI is likely to improve e-governance systems, and more importantly, the quality of statistics: better employment data, for example, helps to craft future economic policies.
Of course, benefits as such are dependent on Nepal’s regulation and the pace of adoption of such technologies. At present, the infrastructures in industries where large gains could be achieved through AI (energy, agriculture) are severely lacking, so steps towards automation and digitization need to be taken. Future regulation should put in perspective the predictions that AI will create massive unemployment worldwide.
The 2016 World Bank Development Report estimates a higher rate of unemployment due to automation in low and middle-income countries than high-income ones. The demand for highly skilled workers is bound to grow as low skilled workers face job losses – this means the state needs to take urgent measures to stop the brain drain by investing more in education and creating an environment for entrepreneurship and businesses to grow. Remittance contribution to the economy will also be affected as demands for low skilled workers decreases abroad. Developing the country’s capital market needs to be a priority to attract foreign direct investment and increase employment locally.
On the other hand, the future looks darker for women as jobs market for Science, Technology, and Engineering, fields dominated by men is bound to experience significant growth. This will be compounded by job losses in office and administrative roles, where a large number of women are employed. Programs by both the private and public sectors to enroll more girls in schools and push them in science and tech careers are required in order to address the upcoming gender disparity in employment.
Besides impacting the economy and employment, AI is optimizing the delivery mechanisms of public goods and services. During the recent earthquake in Nepal, the UN used its 1500 strong group of volunteers of its Digital Humanitarian Network to label the tweets coming out of Nepal as “urgent needs”, “infrastructure damage”, and “response effort”. This data was fed through an AI engine called AIDR (Artificial Intelligence Disaster Response) to generate a “Live Crisis Map”. Flowminder, a Swedish NGO, collaborated with Ncell to use cell phone location data to estimate population movements after the quake in order to inform the relief efforts.
AI was used in Brazil to analyze existing government data along with satellite imagery to predict areas where the greater incidence of Zika could be expected in an upcoming three-month period. AI has also helped to identify crop diseases using image recognition technology via a smartphone. Plantix is an app that can detect more than 240 plant pests and diseases automatically. In Uganda AI has advised optimal routes for emergency vehicles by analyzing road networks.
However, it requires caution when carrying over these technologies to Nepal without accounting for the local realities. For instance, an AI-based traffic system needs to evaluate local driving habits and flow of traffic. Moreover, there have been cases that warn against much reliance on AI. An example is Google Flu Trends that failed to predict flu outbreaks despite having promising beginnings. Importantly, a greater risk from increased automation is the effect on the funding of public services.
The presence of a large cash-based informal economy puts Nepal at a significant risk as unemployment due to automation eats away the tax revenue of the government. Additional funding could also be required for social security support for the new unemployed base, as voices get louder around the world in support of Universal Basic Income (UBS) – a minimum salary distributed by the government for a country’s population. Particular attention also needs to be paid towards the accessibility of AI-based technologies to people from low income and marginalized communities. Economic empowerment of such communities via science and tech education should also be pursued, while their involvement in developing AI technologies could reduce risks of inapplicability as it happens when the systems come from the developed world.
Finally, AI’s role is notable in increasing civic participation in countries where multiple languages are spoken. The 2011 census found that 123 languages are spoken as mother tongue in Nepal. Many communities still can’t converse in Nepali. The same census reports a 65.9% literacy rate, with female literacy at a mere 57.4% as compared to male literacy of 75.1%.
AI-based automated translation and speech recognition systems can allow interaction with public services and the government via spoken means. Fusemachines is a Nepali AI startup working on Nepali Automated Speech Recognition as one of their projects. Inclusiveness of more languages in future could help a considerable percentage of the population. As the internet usage by women is much lower than usage by men, these technologies could make their voices more heard in civic discourse.
On the opposition, AI could be used to identify political opponents, critics, and dissidents. There are also concerns about the distribution of “fake news” and misinformation, particularly during election periods. The results of the recent US election and Brexit referendum were allegedly affected by such activities. As Nepal transitions to a federalist state, it needs to monitor and control such anti-democratic uses of AI systems.
A new set of opportunities and challenges will emerge in a future powered by artificial intelligence. In order to fully leverage the benefits of these new technologies, preparatory changes are required in different industries. Digitization and automation will increase the industrial readiness for adoption of AI-systems.
However, the rise in productivity achieved via these systems also makes it harder for the state to meet employment demands, thus increasing pressures on its coffers. Nepal needs a prudent and methodical approach towards future policy making to obviate a combustible population, dissatisfied by inequality and further marginalization.
(Partly based on the Web Foundation’s report on Artificial Intelligence for low and middle income countries)
This is a guest post by Rubin Ghimire, an entrepreneur and creative mind based in Kathmandu.
Note: The views & opinions expressed in the guest posts are those of the guest author and do not necessarily reflect the opinions & views of TechLekh as a whole.
Do you want to share your story on TechLekh? Publish a piece of writing that can relate to our theme “Tech in Nepal”? If yes, then share your article with us at [email protected], and we promise to publish it.