Artificial intelligence both a boon and a threat
A solution to high drug prices may lie in artificial intelligence. In the past, drugs were created through the tremendously time-consuming process of trial and error.
It is widely estimated that only one out of 8,000 drug compounds tested in labs will eventually reach mass production.
As a result, pharmaceutical companies often pass on their huge research and development costs to consumers in the form of expensive drugs.
This is changing with the implementation of AI in drug research. The introduction of powerful computers and algorithms will likely free chemists from extended weeks and months in labs analyzing molecules and compounds that are bound to fail.
Reducing work hours and R&D costs will lead to lower prices for drugs previously too expensive for many patients.
AI is now being used in concrete ways to provide support in drug discovery, Ronald Vale told the recent World Laureates Forum in Shanghai.
Vale is a fellow of the American Academy of Arts and Sciences who received the Albert Lasker Award in 2012.
In a development that will make life easier for biologists, AI can help rearrange tens of millions of human antibodies in studies on gene transcription, according to scientists at the forum.
Sophisticated algorithms developed through deep learning pave the way for discoveries and creations of new compounds to treat diseases. The prediction by AI guru Martin Ford that robotics will one day supplant humans has long worried mankind.
Mao Junfa, a member of the Chinese Academy of Sciences who teaches at Shanghai Jiao Tong University, is a proponent of adopting AI to sift through literature for clues leading to groundbreaking discoveries.
“We can use big data to find herbal medicines that weren’t discovered before and benefit mankind,” said Mao.
He said the inspiration came from Tu Youyou, China’s Nobel laureate in physiology in 2015. Tu and her team found the chemical substance called artemisinin in the 1970s that played a key role in saving millions suffering from malaria in the developing world.
In addition to drug research, another big user scenario for AI is life science, especially brain and neuroscience.
Neural scientists have identified a human navigation system powered mainly by the hippocampus — a part of brain responsible for spatial memory — and grid cells inside human brains.
The discovery of grid cells enabled Norwegian scientists Edvard Moser and May-Britt Moser to share the Nobel Prize for Physiology in 2014 with American scientist John O’Keefe.
The exponential growth of AI is such that machines can now be made to resemble parts of the human brain in performing cognitive functions like navigation, Edvard Moser said at the recent Shanghai forum.
“If you give a computer information about direction and position, it actually develops a (navigation) solution of its own,” he noted.
Still, the advent of AI is not at a point where we know exactly what are the computations going on in the brain, he said.
Insofar as some are eager to empower machines to think and act like people, “it’s still a long way before we can reproduce emotions in an initial way in computers.”
May-Britt Moser used cloning as a metaphor to explain why it is well-nigh impossible to create an AI-powered machine “brain” exactly like a human’s.
Since human genes constantly interact with the environment, variations resulting from these interactions are unlikely to be replicated or even simulated using AI or cloning, she said.
She cautioned scientists against fiddling with tools just for the sake of efficiency.
“We need science to address problems, but we also have to be aware of the ethical question,” she said.