Hello Mr Walshe and thank you for your input.

Our Research and Development team has already taken steps to participate national standardisation board SFS and in there we can participate to the ISO standardisation. 

I'm very pleased to learn that there's already a group that is preparing ISO standards for an AI. 

Standardisation of an AI is yet very multi faceted task. Given all the concerns that people have and the possibilities that AI has and will have, makes standardisation a tall order. 

Basically a standard is 'an agreement how something should be done', if i'm right? 

Harware level alone may prove to be more convoluted than we would like to think, but there's also aspects of security, traceability and rights for those who are concerned this face of an AI. 

Future is an aspect that we may not even know yet, but still there should be guidelines and standards prepared for it, so that AI would be compatible and safe after new inventions are here.

In short:  AI standardisation have to start with a step backwards in order to see the whole stack of an AI - with everything that is around it as well.  We cannot afford to do just hardware standardisation anymore but we need to evolve even standards with this effort so that we can get it properly done.

I hope that there's a lot of discussion of this matter also here in this forum.  

wbr Kai Salmela  Robocoast team , EU DIH