Brainchip unveils Akida Architecture, page-133

Hi Ryzie

I was reading an interview with a magistrate who works in the far west of NSW. Amongst the many things she spoke about was her observation that she was called upon to deal with large numbers of unlicensed drivers. She said she found this frustrating as she could see no way of reducing the prevalence of the offence because:

1. The local police knew who all the unlicensed drivers were so were incredibly effective in identifying and apprehending them;
2. The repeat offenders were illiterate and incapable of obtaining their licences;
3. To survive they needed to drive because of the remoteness of where they lived; and
4. There is a complete absence of any forms of public transport.

Driving unlicensed is an absolute offence and as such the reality recognised by the magistrate cannot be taken into account by her and so apart from speaking out about the problem the magistrate has no choice but to impose ever increasing penalties as required by the law for repeat offenders.

It seems obvious that an intelligent machine could carry out the role of the magistrate as presently required where unlicensed driving is concerned.

The real issue is how do you teach an intelligent machine to identify 'the grey areas' surrounding most if not all human interactions and then speak out about it as this magistrate is doing.

Extremist ideology has always been a problem for society because there is no tolerance or room for 'the grey areas'.

Unless intelligent machines can recognise 'the grey areas' even what we consider common place moral values and laws no matter how well taught are at risk of being executed in a fashion which we would recognise as extremist ideology.