Why AI struggles in Covid-19?

0
713

While writing this article, the Covid-19 is widely spreading in Russia, Brazil and other South American nations. Within this time period, the world experienced new practices like using masks, sanitizers and were forced to stay in their homes for two months. New terminologies like Social distancing and Work from Home are becoming a part of the daily talks. The world is believing that future technologies will be able to push away the havoc created by the Coronavirus, but there exist challenges for future technologies. Let’s discuss why AI is not able to save us from Covid-19 now.

During the initial days of lockdown itself, authorities warned citizens to keep social distance to avoid the spread of the virus. At that point of time itself, the talks related to AI started. Automation and Robotics were introduced in public places. The move was first started in China, were government deployed autonomous cars and drones for transporting goods and medicines. Later AI apps and systems were introduced in South Korea, US, Japan, India and in other European nations. But the initiative faced so many obstacles. The reason behind is that is Data and Time.

Artificial Intelligence works with data & Covid-19 was a new experience to the humans, so data availability messed up the processes. Only by providing the right data, the AI can function. During these lockdown time, the data availability faced problems and thereby developments in AI faced troubles. Along with that, time is also an important factor. Every development needs time to study the after-effects and its impact on the market. AI also needs time to integrate with the systems.

For example; take the case of drug development, DeepMind is an AI company of Google and the company had an AlphaFold system for developing protein modelling. AlphaFold system of DeepMind is one of the fastest system available in the world for protein modelling. In the lab, it takes months for protein modelling, but DeepMind can do it in days. But the team cautioned that the models creating by the DeepMind may be approximations and may not suitable for final experiments. Plenty of protein targets have been developed in labs and it would be risky to do it by AI during this time of uncertainty because the developers have to spend their precious time to start from scratch.

It’s data that guides AI and we have to gather and organize it in the right way in machines. Healthcare system may not be able to integrate the information easily to the machines because it takes time, experimentations and other privacy regulations.