AI chatbots can give wrong answers. AI systems that control aircraft, vehicles, and robots can pose a completely different kind of problem. As artificial intelligence moves into the physical world, mistakes can ground planes, cause vehicle accidents, and compromise missions. So how do we know when an autonomous system is actually ready to leave the lab?
At TechCrunch Disrupt 2026, Nathan Michael, Chief Technology Officer of Shield AI, Raquel Urtasun, Founder and CEO of Waabi, and Mikell Taylor, Director of Robotics Strategy at General Motors, will take that problem to the real-world AI stage in “Building AI Systems When Failure Isn’t Allowed.”
The conversation will cover the work behind high-risk AI deployments, including building a culture of safety, testing and validating systems, navigating regulatory hurdles, and gaining trust.
Do you want to build AI for the physical world? Tomorrow is the last day to save up to $200 with a Disrupt pass. Secure your ticket by September 25th at 11:59pm PT to hear from three leaders in the field where failure has real-world consequences. Starting September 25th at 11:59 PM PT, prices will increase.
Are you ready for the mission?
At Shield AI, Nathan Michael leads the development and deployment of Hivemind, the company’s platform-independent mission autonomy software. His background spans AI, control, perception, and multirobot systems, including many years at Carnegie Mellon University’s Robotics Institute, where he directed the Resilient Intelligent Systems Lab.
That expertise powers technology designed for some of AI’s most dangerous environments. In February, Hivemind was selected as the autonomous provider for the U.S. Air Force’s joint fighter drone prototype program. A month later, Shield AI announced a $1.5 billion Series G funding, giving it a post-money valuation of $12.7 billion.
At Disrupt, Michael brings the perspective of an AI leader developing autonomous systems for environments where performance must be guaranteed.
What does it take to put AI in charge? Secure your Disrupt pass by September 25th at 11:59pm PT to hear Nathan Michael’s perspective on developing autonomy in high-stakes environments.
Are you ready?
Raquel Urtasun has worked in the field of AI and self-driving cars for 25 years. Prior to founding Waabi, he served as Chief Scientist and Head of Research and Development at Uber ATG. She is also a professor of computer science at the University of Toronto, co-founded the Vector Institute for AI, and has published over 200 AI papers.
Now, Waabi is preparing to further advance its self-driving technology. The company raised $1 billion in January and announced a partnership with Uber to support the deployment of more than 25,000 Waabi Driver-powered robotaxis.
Testing and validation are at the heart of Waabi’s approach. The company’s Waabi World simulator trains, tests and stress-tests Waabi Drivers in a virtual environment, but Urtasun said the company’s self-driving trucks still need to be fully validated before being deployed unmanned.
This puts one of the central questions of the session squarely in Urtasun’s wheelhouse. How do you know when an AI system is ready to operate without a human at the wheel?
Want to learn how leaders in autonomous driving are tackling testing, validation, and deployment? Save up to $200 and get a second pass for 50% off on select ticket types when you secure your Disrupt Pass by September 25th at 11:59 PM PT.
Are you ready to work with people?
Mikel Taylor has spent more than 20 years building robots designed to perform useful tasks. She currently leads the robotics strategy for General Motors’ Autonomous Robotics Center. Previously, he led the Amazon Robotics team that developed Proteus, Amazon’s first autonomous mobile robot.
Her robotics career began in a decidedly non-industrial environment. She once built a senior prom date venue using robots. Since then, Taylor has worked on everything from autonomous underwater vehicles to industrial robotic systems, with a focus on robots that are practical, reliable, and can work effectively around people.
Taylor’s experience will shed light on fundamental challenges facing real-world robotics. This means user experience and adoption must be considered from the beginning, as part of both product design and deployment planning. As AI-powered machines leave controlled demonstrations and enter workplaces where humans must rely on machines, the human element will become important.
How do people collaborate and build trustworthy autonomous systems? Register for Disrupt by September 25th at 11:59pm PT to save up to $200 and hear from leaders tackling these problems in the real world.
Learn how to know when AI is ready with Disrupt
aircraft. Self-driving car. industrial robot. Although the environments may be different, all three focus on the same question. If the consequences of failure are real, what does it take to deploy with confidence?
This session provides an opportunity for founders and technology leaders building autonomous systems to compare how leaders in defense, autonomous driving, and industrial robotics approach decision-making between promising and real-world capable systems.
This session is one of more than 200 sessions across six industry stages, roundtables, and breakout sessions at Disrupt, October 13-15 at Moscone West in San Francisco. More than 10,000 founders, investors, executives, and technology leaders are expected to attend, with more than 250 speakers and more than 300 startups exhibiting. Beyond the agenda, matchmaking, dealmaking, and networking create further opportunities to make connections that move your company forward.
“Almost ready” is not enough when failure is not an option. Last two days to save up to $200 on your Disrupt pass. Get 50% off your second pass on select ticket types when you secure your ticket by September 25th at 11:59pm PT.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
