Assessing the near-ubiquitous use of AI at MIT—whether developing or deploying, expediting existing methods or opening new avenues for discovery—a new MIT Technology Review feature highlights where AI is spurring advances and where its promise is yet unrealized. The article cites two examples from the Koch Institute: Giovanni Traverso’s lab has used AI to design nanoparticles for more efficient delivery of RNA therapies and Regina Barzilay developed AI tools that are now widely used across the pharmaceutical industry.
Yet, Barzilay observes, there are areas where anticipated progress has not yet materialized.
“The reason we cannot cure neurodegenerative diseases like Alzheimer’s or very advanced cancer is because we don’t really understand fully—on the molecular level—the disease itself, the drivers, and how to control it,” she says.
She notes that while AI has yielded some “helper tools,” for biological research, it has not made “a significant transformation” of our understandings of disease mechanisms. Her comments underscore the importance of continued basic research in normal and disease biology, integrating multiple experimental methodologies to generate the knowledge and data needed to fully realize AI's potential for improving human health.