[LINK] Company Pays Up after Robot Breaches First Law

[Kim Holburn]

Some robots can move almost faster than the human eye can see and
certainly faster than humans can get out of the way.  You put them in a
marked out space or a closed off space and make sure humans aren't there
and if humans get in there the robot shuts off (you hope).  It's not
new.  I remember seeing it in a large tape robot in a very big
purpose-built cupboard in 2002.

On 2020/05/12 4:18 pm, Bernard Robertson-Dunn wrote:
> On 12/05/2020 11:52 am, Roger Clarke wrote:
>> [I've no idea what a 'robotic cell' is, nor why a worker was in it]
> 
> https://blog.robotiq.com/what-is-robot-cell-deployment-and-what-it-isnt
> 
> To call it a robot and invoke the mystical First Law of Robotics is the
> usual vendor exaggeration. It's automation.
> 
> Essentially a robotic cell is a single machine that has pre-programmed
> actions. The pre-programming may be complex and responds logically and
> deterministicaly to input data from sensors.
> 
> The failure is either in the safety sensors (assuming there are any) or
> in the logic in the software.
> 
> Whatever the cause, it is a case of human error - someone did not
> account for all possibilities in the software or did not maintain the
> equipment so false data was input. Even then it could be argued that not
> enough redundancy was built in to avoid single points of failure.
>