In the four months since the novel coronavirus was first identified, a myriad of AI tools has popped up promising to fight the disease.
The tech has predicted how the virus will spread, calculated ventilator demand, and searched for symptoms by analyzing X-rays, audio of coughs, body temperatures — and even feces. But most of its applications have been more hype than substance. As Neil Lawrence, the former director of machine learning at Amazon Cambridge, told CNBC:
This [pandemic] is showing what bulls–t most AI hype is. It’s great and it will be useful one day but it’s not surprising in a pandemic that we fall back on tried and tested techniques.
But one use of AI is showing serious promise: Drug discovery.
[Read: Stanford teams up with Fitbit to detect coronavirus through wearables]
Earlier this month, BenevolentAI, a London-based startup worth a cool $1bn, announced that it had uncovered a potential treatment for COVID-19. Yesterday, The New York Times’ Cade Metz revealed how they did it.
AI’s search for treatment
BenevolentAI produces a machine learning platform that analyzes data to find new medical treatments. In January, the company turned its focus to the coronavirus.
Its platform searched scientific literature for an existing medication that could also treat COVID-19. Days later, it pinpointed a particularly promising drug: barcitinib, which was originally developed to treat rheumatoid arthritis.
Pharmaceutical giant Eli Lilly manufactures barcitinib as a pill, which is taken once a day to strengthen the immune system as it battles the condition. But BenevolentAI’s research suggested it could also stop the coronavirus from entering the body.
The discovery shocked scientists at Eli Lilly.
“It was crazy,” Christina Gavegnano, who took part in the work with H.I.V, told the NYT. “We kept asking: ‘Who are these people? Does anyone know them?’”
But doctors who read the research recognized the drug’s potential. Not only does it boost the immune system — it could also prevent infection.
Barcitinb will now enter its first large randomized trial in COVID-19 patients, just a few months after BenevolentAI first began its search for a treatment.
The entire piece is worth a read, as it shows the enormous potential of AI drug discovery — and its limits. Find it here on NYT.
Published May 1, 2020 — 11:43 UTC