Aardvark DailyNew Zealand's longest-running online daily news and commentary publication, now in its 25th year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.
Content copyright © 1995 - 2019 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk
Please visit the sponsor!
A number of luminaries within the tech industries have given warnings about the dangers that AI poses to the future of mankind.
Simply building very powerful computer systems, loading them up with "deep learning" code and letting them lose on large sets of data is, some suggest, a recipe for disaster.
The problem is that as these systems learn, we really don't have any idea exactly what's going on inside them. This was most powerfully highlighted in an experiment conducted by FaceBook where two AI systems were linked and it was observed that they very quickly developed their own "secret" language for communicating. Although researchers could see that they were talking to each other, they had no idea what was being said or what information was being exchanged.
Now yes, that's kind of scary... but AI has even scarier potential right here at home.
Take a browse of this Stuff story.
According to Tori Sullivan, a director of EY Law, the IRD plans to use AI systems to catch businesses and individuals who are avoiding their tax obligations.
There's some interesting information revealed in that piece... such as the fact that the IRD can (and does) access the records of loyalty card schemes and through that it can track your purchases, even if you're paying cash.
It also plans to do some analysis of declared profits and expenses across industries so as to highlight any which stand out due to deviating from the norm.
Now on the face of it this is a good thing. If there's some cafe or restaurant putting its cash straight into the manager's pocket instead of through the books then tax is not going to be paid on that money. However, as the article points out, the IRD thumbs its nose at article 11 of the Universal Declaration of Human Rights which very clearly states:
"Everyone charged with a penal offense has the right to be presumed innocent until proved guilty according to law in a public trial"
That's right, as alluded to in the story, an average cafe may be able to turn 1Kg of coffee beans into 100 cups of coffee and thus generate the revenues expected from those cups. However, if your cafe seeks a point of distinction and makes its coffees a little stronger, you might only get 70 cups (and associated revenue) from a kilo of beans. The IRD's computers might well then flag you as being a potential cheater -- and thus you could find them charging you with avoidance and presenting you with a big bill for what they believe you should have earned.
It is then up to you to prove they're wrong. The presumption is not of innocence but of guilt. So much for your human rights -- unless you're Google, Facebook, Vodafone or Apple because, even though these corporations are not people, their blatant juggling of the figures so as to avoid tax is ignored.
So beware, if you don't want to spend the rest of your life being audited and having to try and prove you're not a tax cheat, you must not fall outside the standard deviation in both your earning and spending patterns -- lest the AI demon strike you down and make your life a misery.
Yes, Musk, Hawkings and the others are right... we should fear AI... but perhaps only when it is backed by laws that violate international treaties on human rights.
Postscript: I don't expect much in the way of comments on today's column. I notice that whenever issues relating to the IRD and tax are the subject, Aardvark readers are generally smart enough to keep their heads down and not dare to question the authority and might of our tax collectors. Fair enough, I'll risk taking a bullet for you though.
Please visit the sponsor!
Have your say in the Aardvark Forums.