akarb - Fotolia

AI COVID-19 tech bolsters social distancing, supply chains

As the world waits for staggered reopenings and a return to normal life, AI-based technology is assisting scientists with diagnosis, detection, research and remote-based workforces.

While we've had many pandemics in the past, none have been so life-changing as the struggle against the latest novel coronavirus, COVID-19. The impacts of the pandemic have significant economic and public health consequences -- including widespread effects on education, e-commerce and global supply chains.

With the world's attention on this virus, artificial intelligence researchers, companies and solution providers of all sorts are looking to apply AI and machine learning to the vast range of challenges that the world faces. Many companies are applying AI capabilities to medical and health needs, while others are applying AI to the ongoing challenges faced in the economy. AI-based COVID-19 solutions are bolstering industries to provide healthcare, enterprise communication and ensure social distancing.

AI helping keep people safe and distant

At this moment, there is no vaccine to combat the COVID-19 virus; the primary way to get control over the spread of the virus is through mitigation and suppression. The most effective treatment so far has been to practice self-isolation and to avoid crowded areas through social distancing. In the case of being tested positive, patients are told to quarantine themselves if they are showing manageable symptoms.

U.S.-based Athena Security is repurposing its security-based imaging solution to the healthcare field by analyzing thermal imagery to detect and track potentially sick patients. The company uses thermal imaging combined with algorithms that analyze the body temperatures of people to flag potentially sick individuals traveling in high traffic areas such as airports, stadiums, train stations and other locations.

Other regions have taken much more intrusive -- some might say draconian -- measures in monitoring and policing communities. Among the solutions employed by China were a surveillance system that used facial recognition and temperature detection for the identification of people who may have a fever. This technology was combined with mobile device tracking data and other information to not only spot those who were potentially sick, but match to facial records databases and indicate everyone who they might have potentially infected. In the Sichuan province, officials used AI-powered smart helmets that could identify people with a fever.

Using data analytics and big data, the Chinese government instigated a program whereby they monitored the risk each individual had of contracting the disease. This identification could be made based on the individual's travel history, time spent in virus hot spots and exposure to people who had already contracted the disease. Based on this, the government assigned codes like red, yellow or green to indicate whether individuals are put in quarantine or advised self-isolation. Across China, drones are also used with thermal imaging to track infected patients, as well as to patrol public spaces for curfew compliance. This social tracking approach will probably become more commonplace as countries look to be more forceful and proactive in keeping infected people home and preventing the spread of disease.

How AI is assisting the fight against COVID-19
AI-based technology is assisting diagnosis, detection, supply chains and telemedicine.

Handling the wave of healthcare and employment claims

When a pandemic hits, no aspect of the global economy is untouched. Health insurance providers and healthcare officials are backlogged by numerous cases of claims that they must process immediately. Likewise, the growing unemployment caused by work closures is resulting in an exponential increase in jobless claim filings. A lot of resources are needed to verify these claims, process them and provide benefits. Furthermore, with government staff themselves working from home and away from internal governmental systems, many of those needed benefits are stuck behind process bottlenecks that require human intervention.

RPA and more cognitive process automation tools that utilize the power of natural language processing for document handling, and more nimble solutions that can dynamically adjust to process changes are being applied to help move claims forward, while minimizing human workload. While RPA adoption has been moving at a fast pace over the past few years, it is expected that the global pandemic and work-from-home requirements will give cognitive automation even more of a push this year.

The growth of video conferencing and chatbots

Likewise, the shift to work-from-home and home education has skyrocketed the demand for online conferencing and education platforms. This has in turn skyrocketed the consumption of the internet and is taxing global broadband providers. While internet providers work to adjust to the new normal of stay-at-home workers, the growth in online platforms is presenting additional opportunities enabled by AI.

As an increasing number of employees work from home, the load on their organization's IT service desks are likewise increasing. Getting employees functional at home is vital to the running of the organization, but this is challenged by the fact that many IT service and operations staff are also working from home. As a result, companies are employing AI-based self-service solutions that can address common and critical IT service needs and resolve them autonomously without human interaction. These intelligent systems can provide step-by-step instructions from IT knowledge bases and the AI-backed digital assistants can help solve these queries freeing up IT for more complex cases.

Routine healthcare has been disrupted by the closure of many traditional doctors' offices, while hospitals must deal with more urgent needs for COVID-19 patients. As such, there's been an increased demand in telemedicine and health-based chatbots that can address a wide range of health concerns. Using these chatbots and intelligent assistants, less face-to-face interaction is needed between patients and medical staff, thereby reducing risks to these individuals. These tools are helping to reduce the overwhelming number of patients that hospitals and medical personnel may face. By employing bots and conversational AI tools it can help assess people with symptoms and address health needs without necessarily requiring an in-person doctor visit. Now, patients that can be managed at home will be advised to stay at home and free up vital resources for more severe cases.

One example of where we are seeing this in action is the Healthcare Bot service by Microsoft that uses AI-enabled chatbots to provide healthcare advice and some telemedicine capabilities. The system uses a natural conversation experience to impart personal health-related information and the government's protocols on dealing with the pandemic.

AI and the supply chain

The demand for online commerce has increased tremendously as people shelter in place. The normal supply routes and logistical supplies suffer as a result of unprecedented lockdowns, closure of nonessential services and even curfews in extreme conditions. One way to address these restrictions is to use technology and robotics driven by AI for the safe provision of supplies, medical drugs and food supplies to those in lockdown.

Terra Drone is providing these services especially in the transportation of high-risk quarantined material and medical samples to and from these sites to Xinchang County's disease control center. This considerably reduces the risk of medical personnel getting harmed by infected people or quarantined stuff. Other companies are utilizing AI to help speed up their logistics and warehouse functions and deliver goods reliably and safely with little disruption to the status quo.

AI seeking cures and treatments

The White House Office of Science and Technology Policy has urged researchers to employ AI to find solutions to issues relating to COVID-19. The U.S. Centers for Disease Control and Prevention (CDC) and World Health Organization (WHO) have likewise asked AI researchers to assist in vaccine research to combat the virus. There are almost 29,000 research documents that need to be analyzed and scrutinized to find information about the novel coronavirus. Computers can extract the required information much faster than humans. To meet this challenge, a Kaggle competition called "CORD-19" was developed to generate potential solutions from interdisciplinary fields to provide input to the available data set as part of this challenge.

One of the most potent capabilities of machine learning is its ability to find patterns in big data. As such, researchers are applying AI in the discovery of potential vaccines and effective treatments. Google subsidiary DeepMind, known for its AlphaZero and research into artificial general intelligence, recently put efforts to find a vaccine through the sequencing of six protein structures linked to COVID-19. Usually, research into vaccines can take a significant amount of time, but using significant GPU-based horsepower and powerful algorithms that can make sense of tremendous amounts of big data, new vaccines could be developed faster using this AI approach.

Companies of all sizes, including startups, are jumping in to help. In February 2020, British startup BenevolentAI published two articles that identified approved drugs that could potentially be used to target and block the viral replication of COVID-19. The AI system mined a large quantity of medical records and identified the patterns and signals which could imply potential solutions. Their system identified a total of six compounds that could block the cellular pathways that allow the virus to replicate. The company is reaching out to potential manufacturers of the identified drugs to pursue clinical trials that can test their efficacy.

Likewise, Insilico Medicine is also applying AI techniques to find a vaccine for COVID-19 and similar viruses. In February 2020, the company generated an extensive list of molecules that could bind a specific protein of the COVID-19 virus. Using their AI-based drug discovery platform utilizing deep learning, the company filtered the potential list of molecules down to just a hundred. They then seek to test seven molecules, which could be put on trial by medical labs for viability as a suitable vaccine for COVID-19.

Other startups such as Gero Pte. Ltd., based in Singapore, are using AI to spot potential anti-COVID-19 compounds that have previously been tested in humans. Using machine learning and AI-based pattern matching, the company identified medicines such as the generic agents niclosamide, used for parasite infections, and nitazoxanide, an antiparasitic and antiviral medicine, that could slow the new virus's replication.

Applying AI to diagnosis and detection

A study published in the journal Radiology wrote that artificial intelligence-based deep learning models can accurately detect COVID-19 and differentiate it from forms of community-acquired pneumonia. The model, which is called the COVID-19 detection neural network (COVNet), extracts visual features from 4,356 computed tomography (CT) exams from 3,322 patients for the detection of COVID-19. To make the model more robust, community-acquired pneumonia (CAP) and non-pneumonia CT exams were included.

With COVID-19 first spreading like wildfire through China, Chinese companies hurried to provide innovative solutions to tackle the problem. Infervision introduced an AI-based solution that uses a machine learning model to increase the speed of medical image analysis and assist with the diagnosis of COVID-19 in patients. The use of AI-enhanced medical imaging reduces time needed to get positive results and can handle the large number of cases that need diagnosis at great speed and efficiency. As a result, hospitals and labs with scarce resources can quickly screen suspected COVID-19 patients and expedite treatment.

In addition to analyzing radiology imagery, AI systems can handle a range of other health-related data and diagnostics. A recent study presented by researchers at the University of Massachusetts Amherst aims to predict illness based on cough patterns. Other AI systems are listening to coughs and can potentially indicate patients who have the coronavirus from other patients who might have coughs originating from other illnesses. The combination of inputs from thermal images and audio input by microphones can assist clinics and other locations in identifying and segregating sick patients.

As can be seen above, the impacts of a global pandemic are widespread, impacting almost every corner of our society and economy. AI is being applied in a widespread manner as well, handling everything from the treatment and prevention of the virus to dealing with the impacts of the pandemic across the ecosystem. No doubt this is AI's moment to shine and show how it can add transformational value across the globe.

Next Steps

Employers face daunting COVID-19 vaccination questions

Dig Deeper on AI technologies

Business Analytics
Data Management