Tag: Ethichsoftechnology

Cathy O’neil explores the ethics behind using algorithms to determine if specific people are adequate matches for certain jobs. Specifically she analyzes tests such as the Kronos test which determines if someone would be a good fit for a job, and the impact tool to weed out underperforming teachers. Both are computerized tests that lack human input, which leaves out many important factors in determining someone’s ability to do a job. Comparing the Kronos test to a human creates two very different processes, I’ll demonstrate through the DIKW framework. One data point the Kronos test may receive is “yes”. The information test has it that this is a response to the test question “Get mad easily?” which the Kronos test knowledge would understand this as a red flag. The wisdom in this situation would be that this person may not be good to hire since they may act irrationally due to their anger. The test doesn’t leave room for explanation or justification as a human might. An interviewer in the same situation may go through a different process. In this situation the interviewer would receive the same data point and information, but the knowledge and wisdom would be different since an interviewer can see things in shades of gray unlike the Kronos test. The knowledge an interviewer may have is “this person may have a mental illness” and then would be followed by the wisdom “I should ask them to elaborate further on why they may get angry easily to better understand the candidate”. There are many contributing factors to why someone responds the way they do on a test, using tests like these are good ways to gather information about a job candidate, but one cannot rely solely upon a test otherwise they won’t get the full picture. The same situation applies in the example including the Impact tool, this only uses mathematical scores, it doesn’t include human responses. In the case with Sarah Wysocki she was known to be a great educator, but her impact score said otherwise and she lost her job. The DIKW framework in this situation would result similarly to that of the Kronos test. The data the algorithm receives is just a number, it reduces a person down to one single score, which isn’t a fair way to estimate someone’s value. The computer would receive this number with the information including many similar scores from other teachers. The knowledge the impact score has is to compare these scores to find the lowest scores. The wisdom that the algorithm shows is that the lower scoring teachers are less fit for their job and therefore should be laid off. This method is unethical and eliminates possibilities for having strong educators like the Kronos test eliminates the possibility of some good workers.