• External Contributor

Dear COVID-19, Here's What AI Has To Say

Hello valued readers, please follow the most scientific and evidence-based suggestions that our government recommends (https://www.gov.uk/coronavirus), where appropriate, I urge you to follow lockdown and maintain social distancing for the good of yourself and families. One instance of contact is all that is necessary for you to contribute to the spread of this virus, even to those whom you love, to those who have worked day and night to become clinical staff, it would be nothing but a great shame to endanger far more lives for your own mistakes - it’s difficult but we will all get through it as a united globe. Stay safe, and best wishes from myself and all at AIDaily.


The modern world has one crucial flaw - mistaking memory for intellect. Intellect is defined as ‘the faculty of reasoning and understanding objectively, especially with regard to abstract matters’. A student with considerable marks in an exam could be considered an intellect, but how often do exams assess how much someone can really understand, as opposed to memorise. It is clear that good memory is a necessary condition of intelligence, but this only gives half the story. A problem has two stages; the first being recalling relevant information, and the second involves applying it to a new scenario. The latter is often difficult to assess, due to the inconsistency of someone’s abstract thinking due to the many factors that influence the workings of the fragile mind. As a result, it is often neglected in assessment but this is a crucial step of being ‘intelligent’, especially as the world tries to solve the truly puzzling circumstance of the novel coronavirus, such strength of mind has not yet been needed so greatly in modern history.

James Flynn, an ‘IQ Expert’ and writer of ‘What Is Intelligence?’

However, there is a form of intelligence that has been built for consistency, to truimph the human brain statistically and aims to improve with each cycle of thought - artifical intelligence. As the details of the Covid-19 virus are discovered (click here for the BMJ’s take - in short, Covid-19 is clearly far beyond the viral pneumonia initally hinted towards, and rather a severely complex clinical syndrome), artifical intelligence hopes to optimise research into the potential applications of pre-existing drugs into curing Covid-19, and to understand the risk factors of the virus, and how to understand the prognosis of patients to help inform clinical decisions and derive pubic information guidelines. PrecisionLife, an AI precision company with a focus on medical healthcare, utlitised their AI-enabled precision platform to conduct a study that uncovered:

  • 70 sepsis related genes, which the study suggests are ‘potentially relevant to severe COVID-19 patients’ and can be thought of as COVID-19 risk factor genes. Understanding of these genes could help to curve the septic pathogenesis observed in these patients.

  • 59 re-purposing drug candidates - i.e. could be used to develop new therapeutic treatments to help students who develop sepsis during severe bouts of Covid-19. These compounds are supposedly able to target 13 of the targets mentioned above.

  • Relevant mechanisms of action include ‘endothelial cell dysfunction, PI3K/mTOR pathway signaling, immune response regulation, aberrant GABA and neurogenic signaling’

This is good news in the treatment of late-stage disease, where the onset of sepsis can often be life-threating. The disovery of relevant pre-existing drugs could offer a crucial but rapid solution in keeping the most severely septic patients alive, delaying the progression of the virus so other therapies can be attempt, or give the body a longer time window to combat the virus by which time recovery could be possible, unless the body succumbs to other symptoms. However, it is hoped that a similar approach could be applied for other common symptons in the progression of the COVID-19 symptoms to reduce the mortality rate of those enduring late stage COVID-19.

Remdesivir, cleared by the FDA to treat the virus, initial studies show that patients recovered faster after administration of the drug.

In addition to this, our beloved DeepMind (who are owned by Alphabet Inc, the parent company of Google) have begun studies into predicting the protein structures of the coronavirus, hoping to fathom how it develops functionally within a victim, how it is to be controlled and could possibly evolve and cause another pandemic episode in the future. In addition to this, it is hoped that artificial intelligence can help doctors understand the biomarkers associated with the progression of the disease, to have assistance in detecting disease markers that help to indicate the stage at which the disease operates within a patient to help assign priority to patients, this sort of assistive role is one that AI is slowing growing into comprehensively across the medical industry. With the government’s testing targets, it is believed that AI could also help speed up testing results and predict when sudden outbreaks are most likely - possibly through using the historical trends associated with similar viruses such as the SARS outbreak.


An incredible ode of thanks to all those at PrecisionLife who conducted the aforementioned study. These efforts will be fundamental in the race to cure COVID-19, and are greatly appreciated by the billions of lives that linger in danger as nature’s best pathogen (that is, a virus) challenges our safety. For many decades from now, these efforts will be congratulated and praised by all.


Thank you all for your time in reading, stay safe, and I hope you all stay happy and occupied in self-isolation!


THIS ARTICLE WAS ORIGINALLY PUBLISHED ON AI DAILY BY JOEL BABY ON MAY 6, 2020.