Recently, the government has introduced a mobile app, Corona Tracer BD, that utilizes bluetooth technology to determine if the user is near another Corona Tracer BD app user
The novel coronavirus (Covid-19) pandemic has been raising fears of an economic meltdown worldwide. Scientists have been using Artificial Intelligence (AI) to trace the outbreak in real-time, and using data analytics to tell in advance where the virus might surface next and to develop an effective response.
This, if done diligently, can save lives.
The first report of a suspected infection in Wuhan was received by the World Health Organisation (WHO) on December 31, 2019. Data scientists using AI systems warned about the spread of Covid-19 beyond China multiple times more than a week before official information about the pandemic was released by international organisations. This prediction was made using natural language processing (NLP) and machine learning.
AI makes higher-order correlations that a human brain is not adequate enough of making. It can associate datasets that a human being would not be able to bind. AI has been used in data analytics to help determine if people have been infected with coronavirus through observation of symptoms of Covid-19, from images of CT scans of lungs, by monitoring changes in body temperature with the help of wearable sensors etc.
AI systems can analyse scans to ease the burden on radiologists and help predict which patients are most likely to need a ventilator. AI applications can be used to deliver medical supplies by robots or drones, disinfect hospitals and search through databases for the effective drugs that might work against Covid-19.
Computational scientists are coupling AI techniques and algorithms with high-performance computing simulations to speed up the ability to screen billions of existing drugs for their interactions with and ability to disrupt generation of the coronavirus proteins. There are many drug-candidate people who would like to screen in a pandemic, and even with the proliferation of cloud platforms and supercomputers, there just would not be enough computing to test them all.
Evaluating a billion small molecules for their ability to bind the coronavirus proteins would take a decade even for the largest of supercomputers. But integrating AI and machine learning into the simulations on the supercomputers allows the programmes to adapt to new information that gets uncovered as they run.
Foreseeing the structure of coronavirus generated by AI has already saved scientists months of research. Bioinformatics, of which AI is an integral part, has provided substantial support in this case by managing to significantly reduce the time required to develop a prototype vaccine that may be tested on humans.
There are thousands of scholarly papers written by scientists and academics on coronavirus. Researchers working on a cure for Covid-19 must have detailed information the earlier research so that they can work faster without making errors and duplications. The problem is that going through all these research papers would take a very long time for a person, whereas an AI-enabled system can help the researchers get an instant overview.
"Covid-19 Open Research Dataset" (CORD-19), a partnership of several institutions including the White House, has been created as a free resource for the global AI community. It consists of tens of thousands of coronavirus-related papers published so far, and is constantly updated with new publications. AI-enabled systems analyse hundreds of millions of pieces of research material to find solutions to problems.
Scientists have developed a series of algorithms in healthcare settings to assess risks of Covid-19, including an algorithm for the primary measures that need to be undertaken for managing contacts of probable or confirmed Covid-19 patients. Some AI applications can also detect misinformation about the contagion by using machine learning techniques for mining data from social media and tracking sensational or alarming keywords.
With the help of AI, security applications can determine and label online sources that are not deemed authentic or trustworthy for fighting against infodemic, the wide spread of misinformation. Facebook, Google, Twitter and TikTok are collaborating with the WHO to verify and mark false information about Covid-19 using AI technologies.
AI can forecast the number of probable new outbreaks by area as well as the types of demography at risk by processing large chunks of unstructured information. It can also assess and suggest optimum strategies for containing the spread of the virus.
AI applications for facial recognition are being used in many countries to track people who are not wearing masks in public. AI-based fever detection systems and processing of data collected on digital platforms and mobile networks to track people's recent movements have helped contain the spread of the virus, although these might have contributed to the concerns regarding invasion of privacy of the citizens or to the arguably draconian enforcement of restraining actions.
The Chinese authorities are using a system using facial recognition technology that scans and takes photographs of people in public places and are deploying drones to patrol and use thermal imaging to track people violating quarantine regulations. The Russian authorities are using automated facial recognition technology to scan closed-circuit camera footage to identify people who have recently arrived from China. The success of these AI applications will not only depend on their technical capacities, but also on how human regulators and AI developers will supervise their implementation pathways in accordance with the established algorithmic standards, legal principles and ethical safeguards.
Some governments are using geo-location data to monitor and contain the spread of the virus. These data generated by mobile Call Data Records (CDRs) can generate important information using AI to understand shifting and movement of citizens. The movements of the mobile users which accounts for a sizable mass of the population can be determined in near real-time with machine learning. The resulting data and trends are extremely valuable for governments to track Covid-19 contagion, issue warnings among the communities that are susceptible, and understand the effect and consequences of measures taken, such as social distancing and quarantine.
The authorities in Bangladesh have been using data collected from the mobile operators to trace the infected population. Recently, the government has introduced a mobile app, namely Corona Tracer BD, that utilises bluetooth technology to determine if the user is near another Corona Tracer BD app user. It will notify the user if there is a risk of Covid-19 infection by checking if the user has been in recent contact with an infected person.
In case of risk, the user can seek medical help at the earliest and go to self-isolation. Apps like this provide a valuable tool for the authorities to track and curb the spread although they cannot cover the entire population because of people who may not have or are not able to use a smartphone, like the elderly. Often these apps are unable to differentiate between people in the same building and those residing in close surroundings.
In many cases, these GPS-enabled data collecting mobile apps continue to run in the background even when the user is not using the device. Some apps can also exchange data with other apps resulting in detailed information. While the WHO praised South Korea's extensive tracking routines, some uses of the data on movements of the citizens by designated local authorities have raised privacy concerns.
Privacy by design seeks to ensure maximum user privacy by data aggregation, or using unnamed or pseudonymous data. The authorities must ensure legal basis of collection and use of various types of data collected. Usage of these tracing technologies and data collection must be proportionate, and should secure the storage, processing and sharing of the collected data.
The authorities have to make sure that the data obtained are fit for the purpose, and no unnecessary data is collected, keeping the public well-informed of full transparency and assuming accountability. Also, the data should be retained only for the duration as it is required to serve the specific purpose for which it was collected.
The Bangladesh government's official portal on Covid-19, corona.gov.bd, incorporates an AI-based chatbot through which people can access free online health consultation services. AI-enabled chatbots have also been essential communication tools to disseminate authentic and verified information.
AI is also used to examine social media posts to gauge how Covid-19, and the changes that it has brought to the general lifestyle, is affecting mental health. Using AI-driven text analysis, systems can inquire about posts that are hashtagged with Covid-related terms, and correlate it with other datasets on relevant factors including the number of cases, deaths, demographics and more to determine the effects of the virus on mental health.
AI and machine learning have become essential tools in tracking and monitoring cases during this pandemic. The fields of pharmacology, medical care, and mobility analysis to contain the outbreak have found a crucial ally in data science to make headway and deliver results.
Using AI-based technologies has sometimes meant balancing the need to defeat coronavirus with the conflicting need to protect people's privacy. Governments must build citizens' trust in technologies used in this fight. With limited skilled human resources, Bangladesh can reap invaluable dividends by investing in AI-based systems that are responsible.
Syed Almas Kabir, is the president of Bangladesh Association of Software and Information Services