Amazon, as a company, has been working on Artificial Intelligence and Machine Learning for a long period of time. While most of the products are not available for the general public, insider projects are being carried out with the power of AI. One of the Amazon departments that used AI extensively was talent acquisition. It developed an AI-based algorithm that would filter the recruitment process by filtering out candidates. Now, however, the same system has been found to have, throughout the period of functioning, favored men over women when it comes to technical jobs. In other words, the AI was not gender-neutral, per se.
The algorithm was expected to help the talent acquisition team in finding out the right candidates. Based on the information provided, the AI would create scores for each of the candidate before suggesting them to various departments and sub-categories. The process was almost similar to what Amazon does in the online store for filtering out products based on customer reviews. The problem here was that the system was not choosing female candidates for jobs in the technical sector, especially development. This trend became very visible by the start of 2015.
As to why the root of the problem was, Amazon says that the AI had taken input from previous ten years of data. It had taken into consideration data patterns in recruitment. Needless to say, it gives an idea that men dominate the tech industry, which resulted in the AI choosing more men over women for technical posts at the firm. Artificial Intelligence of the recruitment system made itself clear that men are more recommended and suitable for technical jobs — as an ignorant human would have done. Now that the issue has come to effect, Amazon has shut down the AI and will be looking for better alternatives.