What To Know
- AI is expected to play a more predominant role in the field of drug discovery and in the pharmaceutical industry Image Credit.
- Chris Meier, formerly of the pharmaceutical industry and now at Boston Consulting Group, notes that AI’s ability to sift through “an ocean of noise” to detect signals could dramatically reduce the time and cost of early-stage drug development.
AI News: For the first time in medical history, a drug entirely designed using artificial intelligence may soon enter a phase 3 clinical trial—the final and most critical step before regulatory approval. Developed by InSilico Medicine, the small molecule named rentosertib targets idiopathic pulmonary fibrosis (IPF), a devastating lung-scarring disease. In June, InSilico reported positive results from a 71-patient trial in China, demonstrating that rentosertib was safe and well-tolerated. The drug’s unique significance lies in the fact that both its biological target and molecular structure were identified using AI tools—a milestone in pharmaceutical research.
https://www.nature.com/articles/s41591-025-03743-2
AI is expected to play a more predominant role in the field of drug discovery and in the pharmaceutical industry
Image Credit: AI-Generated
While headlines frequently tout that AI will revolutionize drug discovery, many experts remain cautious. They argue that AI has yet to deliver breakthrough outcomes in the later and more expensive stages of drug development. InSilico now faces the difficult challenge of proving its candidate can genuinely treat IPF in large-scale phase 3 trials. This AI News report examines both the promise and the pitfalls of this groundbreaking approach.
The Promise of Artificial Intelligence in Pharma
Artificial intelligence, especially machine learning, thrives on analyzing massive datasets, identifying patterns, and making predictions. Chris Meier, formerly of the pharmaceutical industry and now at Boston Consulting Group, notes that AI’s ability to sift through “an ocean of noise” to detect signals could dramatically reduce the time and cost of early-stage drug development. In 2022, Meier led a study which found that AI-derived molecules entering phase 1 trials had an impressive success rate of 80–90%, compared to the historical average of about 66%.
AI systems can rapidly identify which proteins or genes should be targeted, making the initial drug discovery process more efficient and data-driven. Theoretically, this should speed up timelines, reduce costs, and increase the chances of success. But the reality is more nuanced.
Critics Warn of Overselling the Hype
Some experts caution that AI often repurposes known targets rather than truly discovering new ones. Andreas Bender, a professor of machine learning in medicine at Kalifa University in Abu Dhabi, emphasizes that many AI-selected drug targets are not novel. As a result, while the safety of such targets may already be established, the innovation aspect becomes questionable.
Medicinal chemist Derek Lowe also raises concerns. After reviewing 24 AI-discovered drug candidates, he found that most were aimed at already well-understood disease pathways. He warns against the recurring overhype around AI, recalling previous waves of enthusiasm for computational techniques that failed to meet expectations.
AI’s current capabilities shine in predicting protein structures using large curated datasets, like the Protein Data Bank. But in the messier world of patient data and scientific literature—where failures are often not published—machine learning models are less reliable. Without access to high-quality negative results, AI struggles to build balanced and predictive models.
Major Investments Reflect High Hopes
Despite criticism, the pharmaceutical industry continues to bet big on AI. In January 2024, Alphabet’s Isomorphic Labs inked billion-dollar deals with Eli Lilly and Novartis. Meanwhile, Benevolent signed a $594 million agreement with Merck in 2023. Recursion, a pioneer in this space since 2013, has automated labs generating massive biological and chemical datasets to identify promising drug candidates. One of their drugs for solid tumors moved from target discovery to clinical readiness in just 18 months—less than half the industry average.
But even with these advances, challenges remain. AI is excellent at suggesting molecules that target known proteins or genes. However, it still falls short in anticipating complex biological interactions in the human body—such as side effects and organ toxicity—which often become evident only in the later stages of testing.
A Costly Road Ahead
Phase 1 trials typically involve just 15 to 20 people and are relatively inexpensive. But the next stages escalate quickly—phase 2 trials can cost $45 million or more, and phase 3 trials may take several years and hundreds of millions of dollars, with success rates hovering around 55%. When AI-discovered drugs fail at this stage, the financial loss is significant.
This is where InSilico’s rentosertib could change perceptions. If it succeeds, it will be the first AI-designed drug to survive all regulatory hurdles, offering proof that the technology can deliver on its promise.
Cautious Optimism Amid Market Challenges
Not all news has been positive. AI biotech firm Exscientia slashed its workforce and scaled down operations in 2024 before being acquired by Recursion. Germany’s Evotec reduced its pipeline by 30%. Recursion itself ended three clinical programs in May, citing commercial reasons, and is now focused on oncology and rare diseases.
Still, AI is slowly integrating into the drug development mainstream. Even skeptics like Derek Lowe believe in its long-term potential. “I am a short-term pessimist and a long-term optimist,” he says. “I see no reason why these techniques can’t do great things.”
While the full impact of AI on drug development remains to be seen, its growing role is undeniable. The current wave of AI-generated drug candidates may not revolutionize the industry overnight, but they are setting the stage for a future where data-driven discovery and precision medicine become the norm. The tools are evolving, and so is the science.
Drug discovery has always been a high-risk, high-reward endeavor. With AI now taking center stage, the stakes are even higher. If this new generation of AI-powered pharmaceuticals can demonstrate real efficacy in large human trials, it could fundamentally reshape the economics, speed, and success rate of modern medicine. For now, the world watches as rentosertib and its creators prepare to make history.
For the latest on AI in the drug discovery arena, keep on logging to Thailand AI News.
