AI has helped hasten the process of knowledge creation for public good during the COVID-19 pandemic

The ongoing COVID-19 pandemic has clearly highlighted the leverage of AI in the context of information and knowledge for the public good. Notable economists, Nelson (1959) and Kenneth Arrow (1962), described information and knowledge with reference to two related public good characteristics. These public good characteristics are non-rivalrous consumption and non-excludability. Non-rivalry means that once the knowledge is produced it can be consumed by anyone without diminishing the resource for another person.

The non-excludability nature refers to the fact that the cost of excluding one consumer from the use of the knowledge is so high that no profit-maximizing firm is willing to supply the goods. The COVID-19 pandemic has provided ample examples of how non-rivalrous consumption and non-excludability of knowledge has helped avert a greater disaster in a globally interconnected population. AI has helped hasten the process of knowledge creation for the public good. Examples include:

CORD-19

The COVID-19 Open Research Dataset has been chaperoned by the US White House and partners consisting of 200,000+ scholarly articles on COVID-19 and related viruses. The goal is to enable researchers to apply AI and natural language processing to extract meaningful insights into the fight against the pandemic.

DarwinAI

DarwinAI, a Canadian company, with founders who are also researchers at the University of Waterloo, has made public a COVIDx dataset and an AI model, which now has 16,000+ chest X-Rays across 13,000+ patient cases. The data covers bacterial infections, non-COVID-19 viral infections, COVID-19 infections, and normal X-Rays. The researchers claim that experimental results showed that COVID-Net, the AI model, can detect COVID-19 infection with a positive predictive value (PPV) of 88.9 per cent.

Government deployment of AI

The initial set of contact tracing applications by the Governments of India, UK, Ireland, Germany, and the USA have struggled with adoption rates. Reasons for the low adoption rates include issues of trust, high power usage by a constantly on blue tooth, and the quality of the Apps. The future of such Apps as economies open up depend on lower power consumption, open-source for public scrutiny, and more intelligent contact tracing for accuracy. These areas are being addressed through the deployment of AI.

The commercial industry primarily benefits from the private good of knowledge.

Examples include:

Drugs that help build resistance to COVID-19 have been built on a significant amount of existing public and private knowledge while also creating entirely new volumes of knowledge leveraging AI and natural language processing. Drug synthesis AI program called Synthia, as an example, has been leveraged to explore cheaper and faster ways of producing existing drugs that have shown promise in helping COVID-19 patients.

Combinations of AI and Microfluidics have also been used to accelerate new drug discovery for COVID-19. As an example, StoneWise, a startup in Beijing has designed a drug discovery platform integrating artificial intelligence and microfluidics, with computational chemistry, computational biology, pharmacology, and clinical medicine.

Wearables measuring physiological parameters in combination with AI have the potential to screen for early intervention or track the recovery from COVID-19. As an example, Empatica Inc., an MIT spinoff, utilizes non-invasive measurements from medical smartwatches and artificial intelligence to enable continuous and real-time insight into COVID-19 like symptoms.

India as the data capital

From a Gartner hype cycle standpoint, the technologies described here are past the “trough of disillusionment” and are in the process of navigating the “slope of enlightenment.” For AI models to be successful, the first step is having the data. India can be the data capital for the world. The Government of India and public institutions can help by making appropriate anonymized datasets public. The commercial industry, in India, should be agile in adopting AI models for lower-cost drug development as well as new drug discovery.

Over the next decade, AI will augment decision-making capabilities by providing the right knowledge to the right user (people and machines) at the right time for both the public good and private benefit.