AVI POLICY BRIEF ISSUE: 2020, No. 07 | Fighting the COVID-19 Pandemic with AI

Executive Summary

❖ Although AI is not used widespread in Cambodia yet, the purpose of this article is to review its potential role in countries with advanced science and technology capabilities, in order to reflect on strategically establishing the future ecosystem for AI to develop in Cambodia.

❖ AI was used to detect and cure COVID-19 infections, and prevent further expansion of the pandemic. AI accelerates the identification of the genome sequence of the virus, hence enables the design of early test kits. AI pinpoints potential antiviral drugs and molecules that can be candidates for vaccine production. AI also helps in epidemiological studies including the prediction of outbreaks, spread and mortalities during pandemics. AI plays key roles in addressing the consequences of the pandemic: debunking fake news on social media platforms, enabling home online learning or telework.

❖ From these early experiences of AI in mitigating the COVID-19 pandemic, Cambodia may aim at establishing a proper strategy for preparedness and response to future viral outbreaks, while taking advantage of the emerging digital government and the economy.

❖ Policy Options:

• Prepare various laws to guide and regulate the practices of digitalisation of governance and economy.
• Establish an ecosystem and legal framework to regulate the cyberspace and general data protection.
• Invest in digital infrastructures.
• Build digital talent in all sectors including medicine, public health and global health.

History of AI

Diseases have existed in human history from the dawn of time and epidemics have shaped civilisations along the way. Epidemics started when humankind shifted from the hunter- gatherer to agrarian lifestyle in which clusters of settlements permitted the outbreaks of infectious diseases through human-to-human transmission of germs. The discovery of germ theory gave rise to modern hygiene practices, which soon became the early foundation of public health. Currently, as the coronavirus (COVID-19) rapidly spreads globally and public fear becomes uncontrollable, governments and international organisations race against time trying to contain the spread, while efforts are being made to design vaccines and antivirals to protect public health. While fighting the virus, all stakeholders are also addressing the social and economic impact of this epidemic. In this multi-front global fight against the outbreak of a new virus, Artificial Intelligence (AI) technologies seem to occupy a key position in enabling this global battle. Hence, the innovative and ubiquitous role that AI plays in the current epidemic is worth exploring.

As we learn from myths and stories of ancient civilisations, AI is a concept that dates back to the Ancient Greeks, Chinese, Indians and Egyptians. “Thinking machines” were imagined in ancient Egypt 4500 years ago, while in ancient India, automated guardians had been used to protect the relics of Buddha. Around 900 BCE, in China, Yen Shi showed King Mu an automaton that could sing and act. From the 9th to the 13 th century, Arab innovators and engineers created several automata that could play music or serve drinks. During the Renaissance, Leonardo da Vinci designed automata in the form of a mechanical knight, a lion, and an auto-propelled cart.

Fast forwarding to modern times, AI has now emerged as a technology that offers many practical solutions to social and economic challenges. Yet, many applications have yet to be validated. The Corona virus outbreak opens unlimited opportunities for scientists and engineers to test their innovative prototypes, both hardware and software. New approaches popularised in the West by Deep Blue’s mastery of chess and Watson’s knowledge in “Jeopardy!” have since evolved to play a part in many other fields, including the healthcare system. AI-powered tools and models have been helping in the detection and management of diseases such as breast cancer detection, determination of the Chagas vector and the improvement in communication between healthcare professionals and patients.

AI to Detect, Cure and Prevent

AI technology is currently developing at a breakneck speed in various places across the globe. Complex algorithms and software have enabled the extraction of useful medical information from huge amounts of data that are provided by hospitals and relevant institutions. Such evidence permitted the emergence of predictive medicine and telehealth where scientists, medical and health professionals can now make their decisions to prevent and treat diseases.

In the context of COVID-19, AI technology helps in screening, thus the triage of a large population for fever in order to test them for COVID-19 virus. Thermal sensors placed at airports, train stations or other high traffic public places had not been proven effective, but this policy is in place at almost all airports. AI-temperature screening has been tested to increase the effectiveness of detecting people with fevers. In China’s public areas, staff are using handheld temperature-measuring devices to screen clients. More advanced AI tools are used in the capital’s train stations: Megvii installed remote fever detection structures that can measure in a crowd through masks and hats, sending alerts with an accuracy of 0.3 degrees Celsius. They only require one staff on-scene and have a range of 5 meters. Baidu’s thermal sensors have a smaller margin of error (0.05 degrees Celsius) and can also screen people in motion. Sensetime serves the same purpose, with an additional face recognition functionality, which proves to be useful in the case of people wearing masks. Another interesting tool used in Chengdu is the smart helmet worn by officials in the city, allowing them to scan passers-by at a range of 5 meters for fever. An alarm rings when a high temperature has been detected in a person. One concern in a large country like China is that the epidemic itself is slowing down efforts to distribute such devices, as cities’ transport systems are shutting down, which will dampen the efforts aforementioned.

Beyond the triage stage, AI, the Internet of Things (IOT) and their clinical sensors help in guiding remote patient monitoring and data analytics. Besides fever screening, AI analysis of computerised tomography (CT) scans developed by Alibaba allow for distinction between COVID-19-associated pneumonia and other lung infections. AI algorithms that combine clinical data with laboratory and imaging tests provide physicians with unprecedented evidence for a decisive clinical judgment that guarantees the best care for all patients. From the treatment perspective, machine learning (ML) examines correlation between current treatments and patient outcomes by sifting through big data in the search for the optimal treatment options. The possibility of devising treatments by using AI algorithms to identify both the virus’ structure and the human genome opens opportunity for personalised cures. Beijing Genomics Institute and Baidu have been supplying researchers with such tools. BenevolentAI, a company specialised in “using AI for scientific innovation” offers tools to detect the Janus kinase inhibitor Baricitinib, as a potential new COVID-19 treatment.

AI algorithms, data science and cloud computing can accelerate the production of new vaccines. Their combined computational power helps in identifying relevant data patterns and even drug prospects to design new vaccine molecules that will be tested through animal models and initiate clinical trials processes. Giant Chinese tech companies like Alibaba, Baidu, Tencent, Huawei and DiDi invested heavily in the development of treatment and vaccines against COVID-19.

AI to Predict Patterns and Mortality Rates

AI is a powerful tool for epidemiologists and virologists, as it helps predict the spread and mortality of a disease, thus permitting an early warning of its incidence. Data provided by computation complements the work of those experts, but do not substitute for their expertise. A Canadian AI startup company BlueDot predicted the outbreak of a SARS-like epidemic in Wuhan on 31st December 2019. The US Center for Disease Control recognised the infection on 6 th January 2020 and the WHO notified the public on 9 th January 2020. A young doctor from the Wuhan Central Hospital, Dr. Lee Wenliang had warned the authorities about an unusual viral infection on 30th December 2019, but did not to prevail. He later died from this this viral infection. What happened next, is now history.

Metabiota predicts and tracks early epidemics by analysing flight traveler data. These disease tracker systems use Natural Language Processing (NLP) and Machine Learning (ML) to crunch flight data, news reports, social media, climate, local livestock and various other sources of information to obtain their results. Mortality rates are estimated using a Recurrent Neural Network (RNN). Chatbots are helpful assistants to inform the public about the clinical symptoms and various patterns of the epidemics, thus freeing scarce health professionals with special expertise to attend to actual patients.

AI to Debunk Fake News

AI can read a very large volume (billions) of web pages in almost 30 languages. NLP, voice recognition and tailored algorithms may help in selecting relevant topics and patterns of interest for users. How would AI be able to distinguish real news from fake news is indeed challenging as the definition of fake news is highly debatable, especially from the propaganda perspective. After all, AI is a just a technological tool.

During this epidemic, fake news triggered the panic buying of medicines, toilet paper, and even garlic. Detection of discriminative language related to the stigmatisation of racial minorities can help maintain social order through the removal of harmful content, while this regulation may lead to excessive censoring. Facebook and Snopes terminated their partnership, due to a disagreement about their fact-checking process.

AI to Manage Schools and Work Disruption

Hundreds of millions of students are now out of school during the outbreak of COVID-19. Online learning is becoming the solution to adapt to this massive disruption. AI distance- learning and home-based working systems are gradually becoming the norm as both governments and industries take precautions in this epidemic by telling their employees and citizens to remain at home. In East Asia, children have been staying home since the Lunar New Year, yet classes have been ongoing thanks to the help of online tools, as well as television broadcasting classes across the affected regions. Video-conferencing and other tools from companies such as Bytedance and Wechat have facilitated working away from the office. Ironically, the Millennials may have been finally offered their “preferred” classrooms. This forced learning experience may trigger a much-needed reform of the traditional school centred on teachers. COVID-19 may become the best online education reformer of the 21st century!

Similarly, some of the millions of workers currently confined at home may have the unique opportunity to create new types of jobs or innovative business models, while others have to adapt old habits to meet the automation process of Industry 4.0. Indeed, while the bulk of those workers will return to their routine once the epidemic is over, the most creative and risk-taking individuals will build a new generation workforce that will sustain the incoming fourth industrial revolution.

Policy Options

AI technology may seem premature in Cambodia now, but the government is pushing to enable the digital economy in Industry 4.0 and will provide a conducive ecosystem for AI to play a major role in the Cambodian healthcare system. Therefore, Cambodia should strengthen her readiness to fight future pandemics using AI through the establishment of innovative forward- looking policies:

1. 1. Prepare various laws to guide and regulate the practices of digital government and economy.
2. 2. Establish the ecosystem and legal framework to regulate the cyberspace and general data protection.
3. 3. Invest in digital infrastructures.
4. 4. Build digital talent in all sectors including medicine, public health and global health.

The opinions expressed are the author’s own and do not reflect the views of the Asian Vision Institute.