Most technology awards are decided on paper. Submissions are written, slides are polished, and juries assess promise rather than performance. The Robot City Award at the newly launched Davos Tech Summit chose a radically different path and the calibre of the jury it assembled tells you how seriously that choice was made.
Ayisha Piotti, Managing Partner of RegHorizon and a globally recognised voice in AI policy, was invited to serve on that jury - a select panel entrusted with evaluating Europe's first city-scale deployment of Physical AI: more than forty robotic systems operating live across the streets of Davos. It is not an honour extended lightly. Judging machines in the wild, in front of the public, demands jurors who understand not only what technology can do, but what society will accept it doing. Few sit at that intersection of engineering reality and governance judgment as squarely as she does, which is precisely why she was asked.
The jury did not read submissions. It watched the robots work. Evaluation rested on three criteria no pitch deck could simulate: real-world deployment, societal relevance, and public impact.
The Deployment Gap
Physical AI - robotics animated by modern AI systems is moving rapidly from demonstration to deployment: delivery robots, inspection systems, assistive machines, autonomous vehicles. Yet most of what the public hears about these systems comes from controlled environments: the lab, the demo video, the trade-show floor. The gap between a system that performs under controlled conditions and one that performs on a real street with weather, pedestrians, unpredictable obstacles, and watching citizens is precisely where both the engineering and the governance challenges live.
Evaluating robots in a functioning town, rather than on paper, closes that gap. It rewards what actually works, exposes what does not, and does so in public view. For a technology whose social licence will depend on how it behaves among people, that transparency is not incidental. It is the point. And it is why the composition of the jury mattered as much as the format: the evaluators were, in effect, standing in for the public whose streets these machines will one day share.
Evaluation as a Governance Tool
For Ayisha, whose work at RegHorizon centres on turning AI policy debates into workable practice, the assignment carried a lesson far beyond robotics. Regulators worldwide are wrestling with how to assess AI systems - through audits, sandboxes, conformity assessments, and impact evaluations. Most of these mechanisms remain document-based. The Robot City Award's method suggests a powerful complement: structured, observable, real-world evaluation, where societal relevance and public impact are scoring criteria rather than afterthoughts.
Judging technology by how it behaves in the world, not by how it is described, is a principle governance frameworks would do well to absorb. Having one of AI policy's leading practitioners apply that principle first-hand -robot by robot, street by street - is exactly the kind of bridge between regulation and reality the field needs more of.
The Swiss Thread
It is fitting that this experiment happened in Davos - another stop on a road that runs through the AI Policy Summit in Zurich to the Global AI Summit in Geneva in 2027. Switzerland keeps demonstrating that its contribution to global AI debates is practical as much as diplomatic: creating the settings where technology, policy, and the public actually meet. Ayisha's presence on this jury places her, once again, at the centre of that Swiss thread.
Physical AI is leaving the lab. The question that matters now is how it performs when it does, and this jury, in Davos, was among the first anywhere in the world to answer it by direct observation.