AI Companies Should Target These 7 Pandemic Problems

The coronavirus can teach us to use AI in fighting future pandemics.

COVID-19 has uncovered the specific problems of highly contagious infections. By reviewing them one by one, you can develop a targeted approach with technology.

Let’s take a look at how AI is helping solve those problems now.

1. Unpredictability

The main challenge of a pandemic is that its course is hard to predict. AI can help identify early signs and allow for intelligent decision making.

The two AI techniques you should look at are natural language processing (NLP) and machine learning (ML). They enable processing various data from numerous sources such as official statements, online media, demographics, worldwide flight data and more.

Three AI companies, BlueDot, Dataminr and HealthMap, issued a warning nine days before WHO’s COVID-19 statement.

By using similar technology, the epidemic tracker Metabiota accurately predicted the risk of the coronavirus spreading to South Korea, Taiwan, Thailand, and Japan a week prior.

Another task for AI is uncovering the factors behind virus transmission rates. To make this a reality, the ML community Kaggle has encouraged its members to create case-and-fatality prediction models.

2. Human-to-human spread

Fast community spread is what made COVID-19 so hard to slow down. There are many use cases of AI where it is paired with other technologies to help contain outbreaks.

In Bahrain, the government released the BeAware app in which the infected are mandated to register. Others can use it to spot COVID-19 cases nearby and avoid contact.

Reducing interaction at workplaces is another way to contain the spread. Supermarket chains in Saudi Arabia are using an AI-app based on “aisle-mapping” technology. It helps packers locate items ordered online across the store without crossing paths.

In a worst-case scenario, authorities may decide to impose strict limitations. The police in Dubai detect non-permits on the street using an AI-powered camera network. The cameras recognize people by their face, voice, and license plates.

In Europe, businesses and institutions have joined forces to build the COCOVID app. By combining big data, AI, locating and bluetooth, it will help coordinate testing and enable bookings of test slots.

3. Healthcare overload

The coronavirus put a huge pressure on hospitals. This has left many non-pandemic patients waiting for care, sometimes too long. Easing the burden on medics is a job for AI.

Hospital personnel need to wear heavy personal protective equipment (PPE) when handling COVID-19 patients. Taking it on and off takes time and is prone to human error, making way for infection. And wearing PPE throughout the shift is stressful.

Singapore has solved this problem by using BeamPro, a telepresence AI robot, at hospitals. It checks on isolated patients by chatting with them in four languages and even delivers meals and medications to them.

The recent innovation by MIT’s CSAIL, Emerald, can make things easier yet. When wall-mounted in a ward, this module monitors the patient’s movements, sleeping, and breathing. It uses AI to track movements and discern people from one another.

AI chatbots are increasingly being deployed by emergency departments. By using NLP and ML, they can analyze a caller’s words and speech patterns to spot COVID-19.

Aside from emergency, this technology is taking upon many repetitive tasks, such as:

• assisting callers with appointments;
• consulting about symptoms and advising;
• medical coding and billing;
• and prioritizing cases.

Interestingly, some medical centers in the US have updated their illness course–prediction AI systems to forecast COVID-19 aggravations. This helps optimize resource allocation.

4. Inefficient testing and diagnosing

Hospitals often lack resources to identify infection and AI has proven itself invaluable in this regard.

As COVID-19 ravaged in China, Wuhan doctors scanned the lungs of thousands of patients using Axial AI. In just seconds it identifies patients at high risk of developing viral pneumonia
In Britain, the Royal Bolton Hospital is using an algorithm trained on over 2.5 million X-rays of which 500 are confirmed COVID-19 cases.

An interesting approach to diagnosing has been put forward by Professor Cecilia Mascolo. The expert from the University of Cambridge suggests using AI to analyze cough, voice and breathing recordings. A team at the Wadhwani Institute for Artificial Intelligence in Mumbai is already developing a tool for that.

5. Slow treatment development

The coronavirus has shown that the current speed of vaccine and drug development is inadequate. The AI’s ability to process big data fast can change that.

A team from Northwestern University has pioneered an accurate AI tool that rapidly scans a large amount of research data, weeding out the most valuable studies for drug development.

The predictive capabilities of AI can also speed up treatment delivery. By using genome data, Google’s DeepMind predicts protein structures of organisms. This can help scientists figure out which drugs might be useful against novel viruses.

For vaccine development, look at the efforts of OSE Immunotherapeutics and MAbSilico. By using AI, they have been able to accelerate the optimization of epitopes that can induce a robust cell memory immunity.

6. Misinformation

Conspiracy theories (5G masts, Bill Gates — you name it) and ignorance are paving the way to the spread of misinformation. Research has shown that people relying primarily on the news from social media are more likely to go against containment measures.

Although internet companies have been using AI to fight the spread of misleading articles and posts, it hasn’t been very effective yet.

Facebook uses image-scanning algorithms. Once an article has been labeled as misinformation, the algorithm scans the image used in the original post and attaches the fact-checking label to every publication containing a similar image.

Yet there are still groups promoting coronavirus-related conspiracy while Facebook mistakenly removes harmless and even useful content.

Clearly, social media need to refine their approach to anti-fake AI.

7.Psychological effects

Social isolation helps flatten the curve — and begets psychological problems. AI can help us better understand how lockdown affects mental health and even provide primary care.

A Stanford University professor, Johannes Eichstaedt, used an AI algorithm to analyze more than two million tweets hashtagged with coronavirus-related terms. He found out that the pandemic is making people in cities more anxious than rural-area inhabitants.

AI-powered chatbots like Replika can take upon some of the hotline tasks when counselors have too much on their plates. These bots use NLP and ML to conduct a meaningful dialogue, helping the overwhelmed.

Conclusion

COVID-19 has given direction to AI developers working toward solutions for future outbreaks. The next pandemic could be as contagious as the coronavirus and as deadly as the flu of swines.
Intelligent software will help us prepare ourselves.

If you have an idea of an AI tool that can make a difference in handling future outbreaks, develop it with Intetics Inc.

featured image by Freepik.com

https://intetics.com/blog/ai-companies-should-target-these-7-pandemic-problems