Should AI Really Replace Judges?

Original article was published by Carlos Renato Hertel on Artificial Intelligence on Medium


Should AI Really Replace Judges?

The “Judge Judy” TV series shaped the American viewer’s thoughts on what means to be a judge.

In 2016, a BBC’s article asked an apparently simple question: “Could AI replace judges and lawyers?”

Judges have biases and, although they must follow codes and duties, humans are susceptible to corruption.

Tay, who many thought would give us hope, was a sad example on how AI, as a mirror of our inputs, can be as mean and dangerous as humans, their creators.

The current AI models, however, also are not rid of bias, as Microsoft’s Tay shown in the past.

2016, Tay, a Twitter-based robot, received countless troll inputs.

The originally meant-to-be normal, friendly, American teen bot, turned to far-right, and even posted a picture praising Hitler.

That was a slave market in Valongo, Rio de Janeiro.

Before 1889, Brazilian slavery abolition, many judges, righteous man for the time, were slave-owners.

If you go back further, you will come across witch-burner inquisitors, and morals keeps getting worse as you travel to the past.

One close day, 2020 will be the past. Law and morals will evolve, as they have always been, and those AI judges might go wrong.

In the short term, sentences can be made quickly, taxpayer money may be saved. However, in the farther future, morality will evolve, and the ‘judges’ will not.

If AI was created in the XIX century, it would be judging until today through a slavist mentality. Things would not work well.

Hopefully, folks can understand that, so our grand-grandchildren will not live in a distopic tomorrow.