Always beware of mathwashing!
You still don’t know how mathwashing may hurt you? Keep reading, then.
As somebody said about planning in Soviet Union, “if there is maths in this game, everything is serious”.
This concept, or belief, is very popular these days, and is called “mathwashing”:
“The widely held misconception that because math is involved, algorithms are automatically neutral, and therefore companies and organizations can avoid responsibility by hiding behind them”.
The first three things to know about algorithms are
- Algorithm is just a name for “procedure composed of a sequence of completely specified simple steps”, especially when that procedure is programmed in software
- Algorithms are written by people: the more complex they are, the likelier it is that they mirror what people know, and want
- Data is not automatically objective either
Mathwashing has always existed. That famous sentence that B. Disraeli probably never said, that is “Lies, Damned Lies, and Statistics” means just that:
<u><em><strong>CAPTION:</strong>
<a href="https://www.brainyquote.com/quotes/benjamin_disraeli_107958" target="_blank">Mathwashing 1.0, late 19th century.</a>
</em></u>
The immense processing power available these days enables procedures called machine learning, that can amplify mathwashing. When things go wrong, algorithms created or powered by machine learning techniques can amplify current inequalities, and make accountability impossible.
In other words, mathwashing, especially when done through computers, is a perfect example of the slogan of this website: your rights depend on how software is used AROUND you.
At the same time, the sure presence of mathwashing in our society does NOT mean that people are entitled to be ignorant, or conspirationists: “Everyone is entitled to his own opinion, but not his own facts”.
How can we recognize mathwashing?
There are at least two ways (besides quality education, of course):
- what I write here about algorithms, machine learning, numeracy and related topics
- the resources at mathwashing.com, that I also recommend using, from the same source that explains Social Cooling