The racism of Artificial Intelligence

You can see many opinions (especially in the USA) saying that Artificial Intelligence is racist, which is real but the “fault” is not AI, it is the society in which we live.

We live in a society with inequalities of all kinds, racial, economic, etc … and in most cases the inequalities are due to a mixture of things. In the USA in poor neighborhoods there are more blacks and Hispanics than whites, and the same happens in prisons.

AI uses predictive tools based on events that have already happened, to try to guess what may happen in the future.

If there have been more arrests of people of color in a poor neighborhood, the system will predict similar events.

The information we handle is always biased, whatever it is, and is based on what happened before. Predicting what can happen based on past events means keeping the same biases. And let’s not fool ourselves, we are also like that even if we don’t give it much thought.

All societies are increasingly multiracial and more multicultural, but that does not mean that we all behave in the same way and that society is uniform. There are big differences.

All societies are organized by strata, and the main differentiation is economic. Depending on the economic level, so will be the standard of living, education, life expectancy, etc …

The data that AI algorithms obtain is taken from society, and in some way is a reflection of it. If these systems are applied, the differences are maintained, and it is very difficult to introduce forced variants to reduce them.

In some very sensitive cases, such as policing, it may be best to stop using these tools.

And in many others that do not have social connotations, AI will be introduced more and more until we can barely distinguish it.

These are the things of technology.