As countries around the world scrabble for ways to curb the spread of the coronavirus, one South African academic believes that AI should be added to the global arsenal.
I recently asked one of my doctoral students where, in this time of coronavirus, all the outspoken experts in the field of artificial intelligence (AI) had gone.
South Africa started out with three Covid-19 cases, then seven, then 13, then 16, and by 3 April we had 1505 people testing positive. AI can help to predict complex outcomes in many situations, so why then is so little use being made of AI to chart the path of the pandemic?
The student felt there were not enough positive cases in South Africa to build robust coronavirus AI models.
Although AI is not accurately predicting new cases because of a lack of data, it is contributing to other ways of dealing with the pandemic. For example, when the coronavirus emerged in China in December, images showing the use of technology were like something out of a science fiction movie.
In a video that went viral on social media in February, an elderly woman in China gazes up at a drone talking to her. The translation, “Yes Auntie, this drone is speaking to you.” This garners a slightly bemused smile from the woman, before the drone goes on to say, “You should not walk about without wearing a mask. You’d better go home, and don’t forget to wash your hands.” She heeds the warning and rushes back inside.
This is an example of how technology can feed our darkest fears, particularly when it comes to AI and other advances in the fourth industrial revolution (4IR). It sets off warning bells about state surveillance and invasions of privacy.
Globally, there are now over a million reported infections. Although a fair number of people have recovered, there have been more than 50,000 deaths. The combination of government-enforced curfews, total lockdowns, travel restrictions and social-distancing has left economies ravaged. Unemployment numbers are spiralling and businesses are shutting down. There are now predictions of a recession far worse than that sparked by the global financial crisis in 2008-09.
Yet, amid this frenzy, proponents of AI are responding to the challenge in ways that don’t involve getting shouted at by drones.
Before the lockdown in Beijing, subway passengers were screened for symptoms of the virus by AI and temperature scanners. Similarly, algorithms are being deployed to chest X-rays to differentiate between pneumonia, tuberculosis and the coronavirus. An AI programme has been developed that predicts with up to 80% accuracy, which coronavirus patients will develop severe respiratory disease.
This identifies early on which patients will likely need hospital beds and which can be sent home for self-care. There are, of course, limitations to current AI technology which is primarily focused on reading pictorial scans such as ultrasounds, X-rays and CT scans. Yet these have proven to be useful tools in managing treatment.
AI has also been helpful in identifying which treatments might be effective in treating the coronavirus. For instance, AI healthcare startup Deargen, based in South Korea, published a paper based on a deep learning-based model which detects how strongly a molecule of interest will bind to a target protein – in other words, which existing medications would work best.
It found that HIV medication atazanavir is the most likely to bind and block a prominent protein on the outside of the coronavirus molecule. Similarly, companies such as BenevolentAI, based in the United Kingdom, are using AI and available data to search through existing drugs that could be used to treat coronavirus patients until a vaccine becomes available. Vir Biotechnology and Atomwise, startups in the United States, are using algorithms to identify a molecule that could serve as the basis of treatment. As thousands of research papers are published on possible treatments for the virus, AI is being deployed to identify patterns and probable routes. In March, the United States government launched a project in collaboration with some tech giants and academics to make coronavirus research accessible to AI researchers to aid health experts.
Elsewhere, the use of AI in combating the coronavirus has ranged from robotic cleaners spraying disinfectant in segregated wards to AI voice assistants calling people to advise on home quarantine. In China, autonomous robots have reduced infection rates by replacing human cleaners. The AI voice robot is based on big data and can check information such as personal identity, location and health condition. It can then categorise information and produce daily reports which have helped to monitor the spread of the virus.
AI is also being deployed to ensure the safety of healthcare workers. The computer company HP has delivered more than 1,000 3D printed parts to hospitals near their 3D research and design centres. The company has also made its 3D coronavirus models free to download. At the University of Johannesburg, for instance, the UJ Library Makerspace team has used 3D printing to produce surgical face shields to tackle the shortage of personal protective equipment (PPE) for healthcare workers. So far, ten shields have been distributed to campus clinics at the University, while another fifteen will be delivered to Netcare911. The UJ Library Makerspace laboratory, filled with 3D printing and scanning facilities, robots and smart computer technologies, was launched in 2017.
Tools have also been put in place to help with the flow of information. In South Africa, people have flocked to a digital platform called GovChat which connects citizens to government services. The notification service uses mobile phones to alert health authorities and direct people to suitable medical facilities. The data also assists authorities to identify Covid-19 hotspots at any given time to manage resource allocation.
At the same time, the Council for Scientific and Industrial Research (CSIR) and the Department of Science and Innovation developed a dashboard to streamline the data. The information helps the government’s Covid-19 national command council to better manage the pandemic. Teaming up with Telkom and Samsung, the government is effectively able to track and trace people who may have contracted the virus, using data such as geographic information system mapping from their phones.
In recent days, someone who had been in contact with an infected person received an SMS stating that the Department of Health had registered them. The SMS said they could help contain the virus by confirming registration and doing daily check-ins on the Mpilo app, launched by the Gauteng Department of Health in September last year to improve the patient experience at public healthcare facilities.
These are just some of the uses of AI that have been tapped into so far. We are barely out of the first quarter of the year and already so much has been accomplished. Without the rapid spread of information and advice through social media and messaging services, people would not be aware of the extent of the pandemic, or even what steps they should be taking to minimise the risk of infection.
Now, as the fight against the coronavirus rages, AI is becoming an increasingly potent tool in our arsenal. DM