Artificial intelligence (AI) generally receives notice from its usual users - government, large tech organizations and academics. However, researchers and developers are increasingly using AI
and machine learning to track and monitor COVID-19, as well as, to gain a more comprehensive understanding of the virus. These uncertain times call for extraordinary measures, compelling governments and businesses to learn how to manage privacy and security while using AI and data technologies, to reduce the impact of COVID-19 in the upcoming months.
Here are three ways AI architecture is being leveraged to combat the current situation.
Early Warnings and Alerts:
AI services can predict the outbreak of COVID-19 faster and with less human interpretation to identify a significant target. BlueDot,
a Canadian-based AI model is known to out-predict humans in spotting infectious disease outbreaks. According to accounts, BlueDot predicted the outbreak of COVID-19 at the end of 2019
, issuing a warning to its clients on December 31, 2019, before the World Health Organization’s announcement on January 9, 2020. While BlueDot is a powerful AI tool, other AI platforms also pinpoint the virus within hours of their prediction. AI systems help identify outbreaks and threats by connecting the dots, whether that’s a virtual assistant responding to a question or tracking the cases. However, it's important to remember that human interpretation is necessary for optional application.
Identification and Prediction:
AI can eliminate many false paths and be used to track and predict how COVID-19 will spread over time and across the world. AI has previously tracked the spread of viruses, allowing these systems to be re-trained using the data from COVID-19. At Carnegie Mellon University
, algorithms that are trained to predict the seasonal flu are now being re-trained using new data from COVID-19. Due to the lack of historical and unbiased data available, accurate forecasting systems can be negatively affected. However, businesses and other platforms are implementing AI strategies to moderate content in order to reduce the noise and to help acquire more accurate information about the virus. Large social media platforms such as YouTube and Facebook have implemented AI systems to scan for misconstrued information, due to the reduction of human staff resulting from the stay-at-home policies. It is acknowledged that AI must be accompanied by human input for optimal content moderation. AI forecasts of COVID-19 are still a work in progress, however combined with epidemiological models, such as SIR, Susceptible
, can improve tracking and prediction of the virus.
Treatments and Recovery:
Prior to COVID-19, AI was known for its potential to contribute to new drug discovery. As COVID-19 has moved across the globe, many research labs and data centers have indicated that they are using AI to search for treatments and a vaccine against COVID-19. For example, Google’s DeepMind
has predicted, not yet confirmed, the structure of the proteins of the virus, which will be useful in developing a drug. Besides DeepMind, researchers at a Singaporean firm
have been using AI to identify several existing experimental and approved treatments that could potentially be used to treat COVID-19; however, they still await for peer-review. It is not likely that these treatments, will be available in the near future due to extensive medical and scientific checks and approval. Yet the hope is that AI can accelerate the processes of discovery as well as repurpose existing drugs.
By providing AI models with more data, in turn enabling them to be operational, and by using human input to moderate the system, AI can play a significant role in fighting against COVID-19. AI will be useful in biomedical research
, determining whether any of the current drugs have the ability to combat COVID-19. This pandemic is an interdisciplinary problem, but the implementation of AI systems can offer headway on accurate and timely research options for this outbreak.