Everyone is racing to put robots to work. Arms, mobile fleets and now humanoids are going into warehouses, factories and depots faster than ever. Recently, leadership at world's frontier AI labs collectively argued that frontier AI should be paced. That means not stopped, but held to a speed where independent checks can keep up. Physical AI needs the same, and it has a simpler question to answer first:
How do you know a robot has done what it was tasked to do, accurately and completely?
Today the answer usually comes from the robot itself, or from parties who have a vested interest in telling a positive story.
Why the race to deploy robots has opened a trust gap, and why a machine's own report is not evidence.
What independent checking of physical AI could look like, for operators and for government.
A robot's behaviour scored against its own normal, whoever made it.
Open floor for questions from operators, engineers and policy leaders.

CEO, RideScan
.png)
CRO, RideScan · Moderator
Operations, engineering and automation leaders at companies installing and running robots, and anyone in government or standards bodies thinking about how physical AI should be checked.
15 October 2026 · Time to be confirmed · Live online