U-M Launches Institute for AI-Driven Therapeutics Discovery to Tackle Drug Development’s Toughest Problem

Bringing a new drug to patients is one of the most complex challenges in modern science. It can take 10 to 15 years and cost billions of dollars. On top of that, nearly 90% of drug candidates fail during clinical testing stages.

To confront this harsh reality, the University of Michigan has launched the Institute for AI-Driven Therapeutics Discovery (AI-Tx), a new interdisciplinary effort led by the College of Pharmacy that will use artificial intelligence and machine learning to address the root causes of drug development failure.

“The failure rate hasn’t meaningfully changed in decades,” said Duxin Sun, PhD, director of the institute and Associate Dean of Research in the College of Pharmacy. “If we want different outcomes, we need a fundamentally different approach – one that combines advanced AI with deep expertise in how drugs are actually discovered, tested and translated to patients.”

AI-Tx brings together faculty from the College of Pharmacy, Medical School, Life Sciences Institute, School of Information and College of Engineering to create a single collaborative home for AI-enabled therapeutics discovery. 

Rather than optimizing only individual steps in the process, the institute is designed to tackle the most consequential questions in drug development: choosing the right biological targets, designing more effective molecules and predicting how drugs will perform in humans at clinically relevant doses. By embedding AI across these stages, the team aims to shorten timelines, reduce costs and exponentially improve the success rate.

The institute has been several years in the making. After discussions across campus leadership, the initiative received university-level support, including approval for three new faculty hires and significant investment to build a long-term research hub. But Sun emphasizes that no single lab, or even a single institution, can solve this problem alone.

“Drug discovery is too complex for any one group,” he said. “True progress requires biologists, chemists, pharmacists, clinicians, engineers, data scientists and industry partners working side by side. Collaboration isn’t optional, it’s essential.”

In addition to its campus partners, AI-Tx plans to serve as a hub for broader engagement with other universities, pharmaceutical companies and government agencies. The institute will host an international symposium within the next two years and is developing an academia-industry-government consortium to align priorities and accelerate real-world impact.

“Our goal is practical impact,” Sun said. “If we can cut development time in half, reduce costs dramatically and improve the odds that a drug succeeds, that changes what’s possible for patients everywhere.”