Bias In Algorithms

In this unit, I learned about some of the imperfections when it comes to computers, biases. Biases among algorithms are pretty common since they are made by humans, who all have biases. These Algorithms are being used all the time for extremely important things. Things like criminal profiling, deciding who gets a loan and who doesn’t, deciding who should get a job interview and who doesn’t. All of these are extremely important when it comes to people's lives, one mess up in one of these and it could ruin someone's life. And unfortunately, all of these algorithms used in these different fields have biases.

For example, in one of our TED talk videos we had to listen to it was brought up that a person couldn’t get better credit solely due to the fact that people in their neighborhood did not pay their loans on time usually. So, due to a trend in a certain area, everyone in that area gets punished, even the ones that should be considered innocent to the issue. I personally feel like the big issue we should look out for is criminal profiling. Criminal profiling decides who and who is not suspicious or should be expected to commit a crime. And the thing about algorithms is that in order to work is that they need data to work with and learn from. So, if a criminal profiling algorithm is provided with data of previous criminal history in a city, and that data shows that minorities commit the most crime, then the algorithms are more likely to suspect minorities strictly due to the fact that they are a minority. Although I'm sure it does work sometimes, I believe that algorithms should go through extensive testing in order to not have the imperfections that we as humans have.