Deploying AI to battle pandemic

Eduardo Gonzalez, vice president of Skymind Laboratory of Neurobionix Research

-A +A

Businesses are becoming more aware of the importance of artificial intelligence (AI) following the Covid-19 pandemic, says Eduardo Gonzalez, vice president of Skymind Laboratory of Neurobionix Research (Skymind).

He says the pandemic makes businesses realise that they can’t rely on humans alone to acquire and service their customers, and this is where AI comes in. It helps business owners to analyse big data, identify trends and make faster decisions. It also helps companies to acquire and serve their clients with a more personal touch through AI-related solutions such as chatbots.

“With the lockdown and people getting sick, businesses realise a new risk, which is they can’t always rely on humans to do things. Now, they have got to think about what happens if people can’t work or clients can’t come to their office. And what if they don’t have enough brainpower to analyse the data and make the right decisions quickly in a time of crisis? This is where AI becomes essential,” he adds.

So the pandemic has awakened business owners to the need of investing in AI to lower their risk should a similar situation arise. “And it will. I don’t think the outbreak is over and there are probably going to be more lockdowns implemented all over the world.”

Gonzalez adds that AI adoption was already happening across various industry; the pandemic merely accelerated the process. 

Some of the technology it has developed is especially relevant to battling the pandemic itself. For instance, Skymind Neurobionix Research worked together with the China Academic of Sciences and United Imaging Intelligence to develop Axial AI, a computed tomography (CT) scan diagnosis platform that helps medical teams analyse the progression of Covid-19 in patients.

The three entities developed an AI solution three years ago that can diagnose diseases efficiently and accurately using CT scan data. With enough data being fed into the AI software, it can diagnose diseases within seconds with an accuracy rate of 90%, says Gonzalez. 

This software helps radiologists save a significant amount of time. “A lot of people don’t understand how time-consuming it is to analyse CT scans. For instance, there are 300 to 350 CT scan images for the lung of a patient. And it takes a radiologist about ten minutes to analyse them. With the assistance of the AI, a radiologist would probably need only a minute to do the job,” he says. 

Axial AI was initially used to diagnose pneumonia before it was developed for other use cases such as rib fractures, brain haemorrhage and osteoporosis.

More importantly, the AI-powered diagnosis software has made meaningful contributions in helping the Chinese government contain the Covid-19 outbreak in Wuhan. During the peak of the outbreak in the capital of the Hubei province, Axial AI was upgraded and deployed to help hospitals quickly identify many patients who were suspected of Covid-19 infection. 

“Among those hospitals that used Axial AI, 60 of them used it to diagnose Covid-19 patients. Not many people are familiar with Chinese hospitals. But many do know the Huo Shen Shan and Lei Shen Shan hospitals, which are emergency specialty field hospitals constructed in a very short period in response to the pandemic. The two hospitals used Axial AI.

“I don’t know exactly how many patients were diagnosed through Axial, but Huo Shen Shan and Lei Shen Shan have about a thousand beds. Including the rest of the hospitals, you can imagine that Axial was widely used in China,” he says. 

This month, Gonzalez says, Hospital Tunku Azizah Kuala Lumpur is in the process of evaluating Axial AI. 

“The main concern is whether the data collected by Axial AI from Covid-19 patients in China can correlate well enough with the local population. This is a valid concern. If the hospital finds it useful, we will support them to implement it on a larger scale,” he says.  

Skymind is also engaging hospitals in the UK and Japan on Axial AI, he adds.