For the first time, a drug both discovered and designed by an AI system is officially undergoing mid-stage testing in human clinical trials.
The drug called NS018_055 was developed by the biotech firm Insilico Medicine and is designed to treat idiopathic pulmonary fibrosis, a chronic lung disease known to cause often debilitating breathing problems.
The drug was approved for phase 2 trials, meaning that scientists are now testing whether the drug actually works.
According to the folks at Insilico, it could prove to be a consequential step forward for the burgeoning AI-enhanced pharmaceutical field, which could lead to a completely new — and potentially lucrative — era of drug research and development.
"This first drug candidate that's going to Phase 2 is a true highlight of our end-to-end approach to bridge biology and chemistry with deep learning," Insilico CEO Alex Zhavoronko said in a statement. "This is a significant milestone not only for us, but for everyone in the field of AI-accelerated drug discovery."
The executive made some pretty lofty promises about the future of the tech. According to Zhavoronko, this new era in AI-assisted pharmaceuticals will, in theory, be defined by massive increases in the industry's "productivity."
"For Insilico, it is the moment of truth... but it is also a true test for AI and the entire industry should be watching," Zhavoronkov told the Financial Times. "Our company, and it's a big, bold claim, can double the productivity of pretty much every big pharma company."
In other words, if the CEO is to be believed, the time it takes to develop a drug could be halved with the help of AI.
Insilico has been deploying AI models that can quickly and efficiently churn through massive datasets to come up with new molecules that can eventually be turned into a drug.
In short, the company claims to use AI to first identify a potential target for a pharmaceutical, conceptualize said pharmaceutical candidate, and then predict its effectiveness.
The company is using NVIDIA graphics processing units to power its AI, an approach that is seemingly already paying off. According to NVIDIA, coming up with the drug NS018_055 took less than 18 months, a process that usually takes up to six years.
Insilico has used its AI platform to discover 12 pre-clinical drug candidates, three of which have advanced to early clinical trials so far. Of these three, NS018_055 is the first to make it to Phase 2 trials.
"When we first presented our results, people just did not believe that generative AI systems could achieve this level of diversity, novelty and accuracy," Zhavoronkov added in his statement. "Now that we have an entire pipeline of promising drug candidates, people are realizing that this actually works."
Of course, these are pretty ambitious statements considering that the drug is only starting the years-long process of human clinical trials. Only time will tell if the drug is even effective in humans.
After all, as University of Michigan pharmaceutical sciences professor Duxin Sun noted in an essay in The Conversation last year, roughly 90 percent of drugs that go to clinical trials ultimately fail.
And, as the FT points out, there is precedent for failure in the sector, as exemplified by the London-based biotech AI firm Benevolent AI laying off half of its staff last month after its lead drug candidate flopped.
Despite these early warning signs, pharma firms are investing heavily in AI startups like Insilico. And with this clinical trial milestone under the company's belt, coupled with the gold rush nature of the broader AI marketplace, we'll likely see a lot more money flowing in.
"There is no shortage of interest," Eric Topol, founder and director of the Scripps Research Translational Institute and the author of "Deep Medicine," told the FT. "Every major pharma company has invested in partnerships with at least one, if not multiple, AI companies."
More on AI and medicine: Microsoft Alarmed at Doctors Using ChatGPT to Tell Patients Bad News
Share This Article