Wayne State scientists are developing artificial intelligence models to reduce costs in complex drug design, while Stanford research shows ChatGPT-4 achieves a medical diagnostic accuracy of 92% – outperforming human doctors by a score of 74%. This wave of AI breakthroughs is reflected in the financial markets, where health tech stocks rose 12% in 2024 and AI healthcare companies now command up to five times higher valuations than their non-AI counterparts, driving 38% of healthcare venture capital investments attracted.
AI models can reduce drug development costs
A project at Wayne State University aims to use AI and computer modeling to design new drugs, with a particular focus on complex types of drugs that have traditionally been difficult to simulate.
“Most drug design is done with small organic molecules,” Alice Walker, assistant professor of chemistry at Wayne State, said in a Monday (Oct. 28) press release. “The calculation models we have for this are good, but they have difficulty with unusual drug scaffolds such as sugars and fluorescent molecules. Our laboratory will use this grant to develop new computational techniques and apply them to the development of new carbohydrate drugs and fluorescent imaging agents.”
The study, funded by the National Institutes of Health, could lead to breakthrough treatments with fewer side effects. Walker’s team will combine simulations with machine learning to better understand molecular behavior, particularly in developing cancer drug targets.
AI outperforms doctors in medical diagnosis research
A new study from Stanford University shows that ChatGPT-4 achieved high diagnostic accuracy, scoring 92% when analyzing real patient cases – significantly better than traditional and AI-assisted doctors, which scored 74% and 76% respectively.
Doctors with access to ChatGPT showed no significant improvement in their diagnostic reasoning compared to those without AI assistance, although they completed their assessments slightly faster. The findings suggest that medical professionals may need better training to effectively deploy AI tools in clinical settings.
“Our research shows that ChatGPT has potential as a powerful tool in medical diagnostics, so we were surprised to see that its availability to physicians did not significantly improve clinical reasoning,” study co-lead author Ethan Goh, a postdoctoral researcher at Stanford’s School of Medicine. , according to a blog post.
The study, published in JAMA Network Open, emphasizes that AI will not replace doctors, but could enhance their capabilities. In response to these findings, several major medical institutions have launched ARiSE, a bi-coastal network dedicated to evaluating AI applications in healthcare. The study emphasizes the need for better collaboration between doctors and AI and proper integration of AI tools into clinical practice.
The healthcare sector is recovering
Health tech stocks rose 12% in 2024, while private market valuations returned to peak levels, largely driven by investments in artificial intelligence, according to Bessemer Venture Partners’ annual industry report. The sector’s total public market capitalization now stands at $98 billion, spread across 34 companies, including Tempus and Waystar, both of which went public with initial public offerings (IPOs) earlier this year.
Bessemer partners Sofia Guerra and Steve Kraus wrote in their State of Health Tech 2024 report that the industry has shown “remarkable adaptability and strength in the face of ongoing market challenges.”
AI-focused companies now attract 38% of healthcare venture capital dollars, driving valuations two to five times higher than their non-AI counterparts. However, challenges remain in early-stage financing, with Series A companies taking 50% longer to raise capital compared to previous years.
