François Chollet @fchollet Deep learning @google. Creator of Keras, neural networks library. Author of 'Deep Learning with Python'. Opinions are my own. Jan. 23, 2019 1 min read

"Bias laundering" happens when we choose to ignore the biases of automated decision systems because of the illusion that all algorithms must be objective since they're "driven by math" or "run by a computer".

Algorithmic biases could be hard-coded by the implementer, or could come from a biased choice of features, or could come from biased data (all data being biased in some way), or could simply arise from spurious correlations (overfitting). Math/computers are a detail in the story.

In general, automated decision systems tend to inherit the biases of the human-driven process that they replace. Unfortunately, these biases start to acquire a veneer of objectivity, and become harder to inspect, or fix.

With humans at least, new generations bring change. Algorithmic bias may prove to be more entrenched than human-driven bias, due to the greater indirection and continuity brought by datasets and algorithms, as opposed to someone's judgment...

You can follow @fchollet.


Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.