Eric Horvitz, Ph.D. ’91 M.D. ’94 and Chief Scientific Officer at Microsoft, delivered a forward-looking address at Stanford Graduate School of Business as part of its applied AI initiative, warning that the current era of artificial intelligence represents more than a technological evolution—it could redefine the trajectory of human existence. Speaking with GSB Dean Sarah Soule, Horvitz emphasized that while AI is advancing rapidly, society’s ability to adapt is lagging, creating what he described as an “impedance mismatch” between innovation and integration. Drawing parallels to transformative historical shifts like the advent of steam power and electricity, Horvitz noted that the full societal impact of AI will unfold over decades, not years. He cautioned that the current moment may be seen in hindsight as a pivotal starting point, even as profound changes continue well into the future. A central theme of the discussion was the erosion of truth in the age of deepfakes and synthetic media. Horvitz, a longtime advocate for trustworthy AI, has worked on technologies that embed cryptographic “wax seals” in digital content to verify authenticity. Yet he acknowledged that technical verification alone is insufficient. “We have to red team it and attack it,” he said, referencing a recent Microsoft study that explored how verification tools themselves could be exploited to undermine trust in legitimate content. Despite these challenges, Horvitz expressed strong optimism about AI’s potential in the biosciences. He predicted that within the lifetime of today’s generation, AI could lead to breakthroughs in treating or managing neurodegenerative diseases such as Alzheimer’s and ALS. He also highlighted promising applications in designing new proteins for cancer therapy and advancing biomedical research. Addressing concerns about the role of human mentorship in an automated world, Horvitz reassured students and faculty that human connection remains irreplaceable. “I will always value mentoring others,” he said, arguing that as AI takes over routine cognitive tasks, the importance of human craftsmanship, empathy, and care will only grow. During the Q&A, student Serena M. Lee M.S. ’25 asked about the most pressing open questions in AI safety. Horvitz responded that responsibility for safety is no longer confined to AI developers. He compared leading AI companies to electric power utilities—essential infrastructure providers whose systems must be safe, even when used in ways they cannot fully control. On healthcare, Horvitz stressed the need for AI models that perform reliably across diverse populations and institutions. He pointed out that models trained at one hospital often fail when deployed elsewhere due to differences in data, demographics, and clinical practices. “Society won’t accept safety issues at the edges,” he said, even if average performance improves. For attendees, the talk offered a balanced, hopeful counterpoint to the often alarmist narratives surrounding AI. Simran Mohnani M.B.A. ’26 appreciated the focus on human-centered applications, particularly in medicine and protein design. “That’s what gives me hope,” she said. She also valued the shift away from fear-driven discourse, calling the event a refreshing and constructive alternative to the dominant tone in Silicon Valley.