On Sunday, 4 August 2019 11:40:47 AEST Karl Auer wrote:

> It is now. That doesn't mean it can never be. We should not discard a 
> technology because it is rough now. But likewise we shouldn't trust it while 
> it is still rough - or rather we should know its limits and use it within 
> them.

The author suggests  "Meanwhile, I have been thinking the opposite: that tech 
really has to get a lot better, so that voice commands, or even thought 
commands, can override our inherent bent for sloppiness."

That mis-identifies the problem and proposes to fix it with far more complex 
technology which will probably only make matters worse.  The root issue IMHO 
lies in the design of systems where it's possible for minor and eminently 
predictable errors to have major consequences.

In the case of the unfortunate man arrested for alleged child pornography 
offences, why do police procedures not involve an independent cross-check when 
a suspect is identified?  It must also have been obvious from the outset that 
he was an unlikely suspect.

In the case of the misplaced decimal point (15 degrees 19.8 minutes east 
entered into aircraft navigation computer instead of 151 degrees 9.8 minutes 
east), surely the system could respond by identifying the location of those 
coordinates ("Melbourne") when the error would be obvious.  (I agree there's a 
problem with way-points, but I'm not the relevant expert.)  I think some 
airlines now require these settings to be cross-checked by another member of 
crew.

The case of the hospital drug vending machine is tragic beyond words.  Hospital 
policy should require the identification and dosage of all drugs to be 
cross-checked by two qualified personnel, and I believe that's standard 
practice in Australia.

David L.

_______________________________________________
Link mailing list
Link@mailman.anu.edu.au
http://mailman.anu.edu.au/mailman/listinfo/link

Reply via email to