The ETH AI Center shared details on LinkedIn of recent discussions held with Anthropic researchers, including Dr. Neil Houlsby, Dr. Bobby He, and Dr. Sotiris Anagnostidis.
The center is one of the world’s largest AI research hubs, bringing together more than 1,500 researchers across 16 departments at ETH Zurich on the fundamentals, applications, and impact of AI.
The university said the session considered how research groups can continue to be at the forefront of AI, how safety is integrated into capacity development, and how academic and industry groups can complement each other. The panel was moderated by postdoctoral researcher Frederike Lübeck and included core faculty member Professor Thomas Hoffmann and researcher Dr. Imanol Schlag.
Important messages about cutting-edge research and safety
The discussion emphasized that advances in frontier AI require theoretical understanding as well as empirical advances. “Frontier AI requires not only scale but also theory,” the center said, emphasizing the importance of scientific principles to guide long-term development.
They also emphasized that competency and safety research are now seen as interdependent rather than separate components. Curiosity-driven inquiry is also identified as core to research behavior and is described in terms of “pulling the threads of what we don’t understand.” The center pointed to the importance of academia in basic science, critical thinking, and training the next generation of AI researchers.
The university said it has introduced the Claude Credit program to its research community of more than 1,000 members, accessible through internal channels. Approximately 150 people, including faculty, doctoral students, and visiting researchers, participated in this event.
The Center also mentioned its next Academic Talk Series session scheduled for Thursday, November 27th, focusing on Machine Unlearning under the title. Machine untraining: new settings and algorithms. The talk is part of the ETH AI Center Academic Talk Series (AICATS), which the center describes as a face-to-face forum for research exchange.
“Strengthening the links between ETH Zurich and leading AI research institutes is key to developing safe, competent and scientifically-based AI,” the university concluded.
