As artificial intelligence transitions from digital assistants to controlling physical machinery, the consequences of error shift from mere inconvenience to catastrophic failure. At TechCrunch Disrupt 2026, experts from Shield AI, Waabi, and General Motors will examine the rigorous standards required to deploy autonomous systems in defense, transportation, and industrial sectors.
Nathan Michael, CTO of Shield AI; Raquel Urtasun, founder and CEO of Waabi; and Mikell Taylor, Director of Robotics Strategy at General Motors, are scheduled to speak on the Real World AI Stage during a session titled “Building AI Systems When Failure Is Not an Option.” The discussion will explore the creation of safety cultures, validation methodologies, regulatory navigation, and the processes necessary to earn public trust.
Michael brings extensive experience in AI, control systems, and multi-robot platforms, having previously directed the Resilient Intelligent Systems Lab at Carnegie Mellon University’s Robotics Institute. He currently leads the development of Hivemind, a platform-agnostic mission autonomy software. His work has recently secured a role as an autonomy provider for the U.S. Air Force’s Collaborative Combat Aircraft drone prototype program. Following this selection, Shield AI announced a $1.5 billion Series G funding round, valuing the company at $12.7 billion post-money.
Urtasun offers 25 years of experience in AI and autonomous vehicles, including her tenure as chief scientist and head of R&D at Uber ATG before founding Waabi. She is also a professor at the University of Toronto and a co-founder of the Vector Institute for AI. Under her leadership, Waabi recently raised $1 billion and partnered with Uber to deploy thousands of robotaxis. The company utilizes its Waabi World simulator for rigorous virtual testing, though Urtasun notes that full validation is still required before driverless truck deployment.
Taylor, who previously led the Amazon Robotics team responsible for the Proteus autonomous mobile robot, focuses on creating reliable robots that operate safely alongside humans. Her background includes work on autonomous underwater vehicles and industrial robotics, emphasizing that user experience and adoption must be integrated into product design from the outset.
The session aims to compare how leaders in defense, autonomous driving, and industrial robotics approach the critical decisions that separate experimental technology from field-ready systems. TechCrunch Disrupt takes place October 13–15 at Moscone West in San Francisco, featuring over 200 sessions, 250 speakers, and 300 exhibiting startups.
Attendees can save up to $200 on event passes by purchasing before September 25 at 11:59 p.m. PT. Discounts include a second pass for 50% off select ticket types.
Why are we not talking more about liability? Who pays when a GM truck fails in traffic?
Mikell’s background at Amazon Robotics makes her the most grounded speaker for practical industrial applications here.
Still skeptical about Waabi’s simulator. Virtual tests can never fully replace real-world chaos on highways.
The $1.5 billion funding round is wild. Defense autonomy really moved from theoretical to profitable this decade.