CZI’s Priscilla Chan on ‘Virtual Cells’ AI Models to Cure Diseases

The head of the Chan Zuckerberg Initiative (CZI) believes that artificial intelligence (AI) can accelerate medical research such that it would be possible within 10 to 20 years to stop a disease very early on in its tracks.

AI can “put us in a world where we aren’t just trying to treat disease when it’s out of control. We’re actually preventing it at the earliest stages,” said Priscilla Chan, who leads the philanthropic organization founded by her husband, Meta CEO Mark Zuckerberg, and herself. She detailed that vision during a presentation Thursday (March 13) at SXSW 2025.

The mechanism for this future is a “virtual cell” model that could revolutionize how scientists understand human biology and develop treatments for all diseases. It is an AI model that harnesses generative AI not trained on text but on biological data.

“What if you showed an AI model three images of the human heart at an atomic level? What if you showed it videos of individual cells interacting with other cells? What if you taught it the molecular code and all the structures inside every one of your cells? What you would get is a powerful simulation of how the human cell works,” she said during a firechat chat at SXSW 2025 in Austin, Texas.

“We call that the virtual cell model. And we think a virtual cell model would completely change the way we understand health and medicine,” she said.

CZI is the recipient of Chan and Zuckerberg’s pledge to give away 99% of their Facebook shares over their lifetimes. It focuses on science, education and community building.

In February, the initiative launched the Billion Cells Project, aiming to generate an unprecedented one billion cell dataset. This project expands upon their previous work and plans to use a large-scale cluster of more than 1,000 GPUs to train AI models for advancing biological research.

Read more: Google’s AI ‘Co-Scientist’ Helps Unearth Research Ideas

Chan said the virtual cell model could significantly speed up the pace of medical discoveries by enabling researchers to test hypotheses computationally before moving to laboratory experiments. This could transform drug development, potentially reducing the process from years to days and making treatments for rare diseases more affordable.

“Right now, the drug discovery process is incredibly difficult. It usually takes decades of trial and error and billions and billions of dollars of investment,” Chan said. “A virtual cell changes that equation. Instead of testing candidate molecules one by one in the lab, you can model the disease in software. You can test a million potential therapies.”

Chan said scientists currently don’t know a lot about how cells work. “What does each gene actually do, and what happens when there’s errors in them? And how do those trillions of cell types come from a single fertilized egg to make each one of your unique cells?” she said.

“Those are the deepest, oldest questions in human biology, and if a virtual cell can help us answer those questions, we will be so much closer to actually cure and preventing heart disease, neurodegeneration, cancer, maybe all disease,” Chan added.

She sketched out three scenarios of how life would change if every scientist and doctor had access to the virtual cell model.

• Personalized medicine

Currently, medical researchers study a small number of samples and extrapolate the treatment to the general population. Using the virtual cell, physicians can predict what diseases each patient is susceptible to and how individuals respond to different treatments.

• Faster drug discovery

The drug discovery process typically takes decades of trial and error and billions of dollars. Successful drugs can be derailed by side effects, and treatments for rare diseases are not viable economically because the pool is too small. A virtual cell can test a million potential therapies and screen out drugs that don’t work — leaving promising candidates.

• Help the immune system

The concept behind a large language model can help develop a virtual cell that can detect if there are signs of tumors. Today, by the time most people develop symptoms, the disease has taken hold and the damage is can be irreversible.

The post CZI’s Priscilla Chan on ‘Virtual Cells’ AI Models to Cure Diseases appeared first on PYMNTS.com.