12/12/25

2025 SODP Symposium | Meghana Keshavan - What Journalists Saw First: Reporting Finds AI's Signals

In this session, STAT News journalist Meghana Keshavan, BS discusses how reporting identifies early signals of AI-linked psychosis, emotional destabilization, and compulsive engagement. She highlights the widening gap between real-world lived experience and formal evidence, arguing that journalism’s role is to surface hypotheses that guide clinicians toward the right questions for future research.

About Meghana Keshavan:

Meghana Keshavan is a journalist at STAT News, the health care arm of the Boston Globe. She has been covering the biotech sector for more than a decade, reporting on the intersection of biotechnology, artificial intelligence, drug development, and federal health policy. She writes STAT’s Readout newsletter, a daily compendium of the most important technological news in health care — including artificial intelligence. Her recent coverage includes FDA regulatory processes, advances in mental health treatments, psychedelic-assisted therapy research, and issues affecting the U.S. biopharmaceutical industry. Meghana joined STAT in 2016. Before that, she worked for news outlets like Thomson Reuters, the Detroit Free Press, the Detroit Metro Times and NPR affiliate WDET. As a young lass, Meghana toiled five years as a research peon in a schizophrenia genetics lab. She once spilled ethidium bromide on herself, and might be a mutant.

About the 2025 Society of Digital Psychiatry Symposium:

The 2025 Society of Digital Psychiatry Symposium is an event focused on the intersection of Artificial Intelligence and mental healthcare. With this year’s theme, "Advancing Digital Mental Health Through AI," the symposium provides a unique platform for thought leaders and innovators to explore the latest AI advancements reshaping the mental health landscape.

Previous

2025 SODP Symposium | Steve Siddals - Real-Life Experiences of Generative AI for Mental Health

Next

2025 SODP Symposium | Amir Rahmani - Wearables & LLMs: Agentic AI for Personalized Healthcare