AI Reshapes Drug Development Workflows, But Breakthrough Molecules Still Elusive
Drug companies have announced a slew of deals for tools to unleash the promise of AI, seen as the biggest technological upheaval since the internet, much as other industries are doing.
New Delhi: Artificial intelligence has yet to deliver on the most challenging aspect of drug development, finding new molecules that lead to major medical advances, but it is already streamlining less glamorous parts of the process, industry executives say.
AI is helping find participants and sites for clinical trials and drafting documents for regulators, shaving weeks off these labor-intensive processes, seven large drugmakers and six smaller biotech companies said at the recent JP Morgan Healthcare Conference.
It can take a decade and $2 billion to bring a new drug to market, pharmaceutical companies say. Many, including Eli Lilly , which has partnered with leading chipmaker Nvidia , are betting AI can also improve the success rate of new drugs.
Drug companies have announced a slew of deals for tools to unleash the promise of AI, seen as the biggest technological upheaval since the internet, much as other industries are doing.
Agentic AI - or AI that is autonomous requiring little human intervention - could increase clinical development productivity by about 35% to 45% over the next five years, consultancy McKinsey predicted last year.
Israel-based Teva Pharmaceutical Industries said it is utilizing AI in multiple ways so it can focus on the big picture goal of successfully bringing new drugs to market.
"Everything else that's around that needs to be as efficient and as small as possible," Teva CEO Richard Francis said. "This is where I think AI digitization, modernization, process improvement, all the unsexy stuff that we get actually quite excited about, makes a difference."
THOUSANDS OF DOCUMENTS
Executives from global pharmaceutical companies AstraZeneca , Roche and Pfizer, as well as smaller biotechs like Spyre and Nuvalent, described tracking thousands of pages of documents for regulators, including clinical, safety and manufacturing records.
The documents must be compiled, cross-checked and kept consistent across geographies, often requiring the costly use of outside contractors, AstraZeneca Chief Financial Officer Aradhana Sarin explained.
Jorge Conde, a general partner at venture capital firm Andreessen Horowitz, is investing in fixes to what he calls drug development's "messy middle," including putting $4.3 million into startup Alleviate Health.
He described trial enrollment as a "leaky funnel" in which participants drop out along the way, and sees Alleviate using AI technology to help with patient outreach, education, screening and scheduling.
TD Cowen analyst Brendan Smith said the use of AI, including large language models like Microsoft Copilot, for administrative tasks has become fairly common in the pharmaceutical industry.
But it may take another one to three years before investors can measure how AI has begun speeding up drug development, he said. Quantifying savings is difficult, Smith said, as it depends on how and where the tools are deployed.
NOVARTIS USES AI FOR HEART DRUG
Swiss drugmaker Novartis turned to AI in 2023 when it was starting a 14,000-person, late-stage cardiovascular outcomes trial for its cholesterol drug Leqvio, Chief Medical Officer Shreeram Aradhye said.
The typical four- to six-week site selection process became a two-hour meeting, as AI helped identify higher-performing sites and allowed Novartis to close participant enrollment with only 13 patients above its trial target, Aradhye said.
"AI becomes augmenting intelligence, not artificial intelligence," he added.
A Novartis spokesperson said the time savings afforded by AI can add up to months over a drug-development program.
GSK, GENMAB, ITM PLAN FOR SAVINGS
British drugmaker GSK said it's using a mix of digital and AI tools to reduce manual data collection and aggregation and trial enrollment, and aims to speed up all clinical trials by 15% this way.
That helped save about 8 million pounds ($10.87 million) in costs for late-stage studies of asthma drug Exdensur last year, according to a GSK spokesperson. The drug won U.S. approval last month.
Danish antibody developer Genmab said it plans to deploy Anthropic's Claude chatbot-powered agentic AI to support clinical development priorities.
Hisham Hamadeh, Genmab's head of AI, said the goal is to automate post-trial work, including analysis of data and its transformation into graphs, tables, figures and clinical study reports.
German radiopharmaceuticals firm ITM told Reuters it has figured out how to use AI to convert long trial reports into U.S. FDA template formats, potentially saving weeks of effort requiring several staff, but has yet to deploy it.
Amgen research chief Jay Bradner said AI is delivering on multiple fronts in the drug development and regulatory document preparation process.
"What everybody's waiting for is the AI drug. When do I get the AI drug?" he said. "I actually think those molecules are in pipelines right now."
