And that’s a wrap on an incredible United Nations AI for Good Summit!
As autonomous robots transition out of the laboratory and into our streets, factories, and homes, trustworthy AI is no longer just a digital abstraction—it is a physical necessity. We need a fundamental shift from robots that simply “work” to robot teams that are provably safe, scalable, and accountable to the public.
During my talk, “Trustworthy Robot Teams,” I discussed how we can build this technical and ethical framework. We explored mathematical safety guarantees in dynamic environments, how to scale collision-free coordination for massive collectives by merging traditional computing logic with modern AI, and a solution to the black box problem of AI behavior: motion watermarking.
To prove that accountability can be put directly into the hands of the public, we ran a live demo on stage: “The Digital Fingerprint.” By embedding a secret, invisible digital signature directly into a robot’s movements, we showed how a standard camera or smartphone can analyze its motion in real-time to verify its identity, origin, and underlying AI model, even though the movements look perfectly natural to the naked eye.
The on-site Prorok Lab team (Michael Amir, Manon Flageat, Mateusz Sypniewski) also had a great time connecting with attendees at our stand in the exhibition hall, where we were able to demonstrate this technology in action across both legged and wheeled robots.
Thank you to everyone who attended the talk, visited our booth, and engaged in such thoughtful conversations about the future of responsible AI! Check out some photos from the talk and the exhibition floor below! 👇
#AIforGood #TrustworthyAI #Robotics #MotionWatermarking #ProrokLab #TechForGood #MultiAgentSystems #AI