Problems in machine learning can occur even for companies with years of experience with Artificial Intelligence, Like Amazon.com. The online shopping giant had an internal project that was using Artificial Intelligence to get job applications from people. But it constantly downgraded female applicants.
Sometimes A.I programs try to make decisions by looking at the past record. This Amazon’s program downgraded every applicant which attended all women’s college or had the word “women’s” in their CVs/Resumes. In this case, that bias was the male-dominated working environment.
An unnamed source told The Reuters ;
They literally wanted it to be an engine where I’m going to give you 100 resumes, it will spit out the top five, and we’ll hire those,
When Amazon got to know that they are not getting results based on neutral gender reform and also they couldn’t be sure other biases had not got into this program. So they just decided to remove it entirely last year.
Reuters report says that this program was only for test. They didn’t use it to hire candidates for real jobs at Amazon. They just did for it a test.
Amazon is reportedly going to give another try to an A.I recruitment tool this time “with a focus on diversity”. Many companies are working on A.I recruitment tool because according to them the preferences for candidates can be coded into the program.