This week I made a short report on the Robodebt scheme. The report covered the algorithm itself, its impacts, and the ethics and morals surrounding it. I found this report very interesting and went 400 words over the limit. I probably could have reduced the report down to 250 words, but I did not want to take out content - it was such a big issue and I had to wrap my head around it.
I find the incompetency of the government IT team baffling. The main issue regarding the algorithm was that they used average income data and compared it to welfare data as if it was actual income data. Mistaking averages for real data is something I have done in year 9 statistics, but should not be done in any governmental department. Surely someone thought that this was a bad idea, I like to think I could make a better algorithm.
Ethics are a difficult area to discuss. I had trouble with this because no one knows how competent the government IT team actually is, maybe they just weren’t trying when making the algorithm. With this knowledge, the ethics and morals would be clear - there would be someone to blame, and the ethics of that person would be unjust. Maybe I am approaching this from the wrong angle, it might be better to discuss the ethics of the algorithm as a whole, and how the failure has affected that. Taking welfare money away from people seems immoral to begin with, let alone taking it away from people who did not receive an excess.
Considering the amount of research I put into this task, I wonder if it would be possible for Robodebt to be the topic of my data science report. If so, I would have already completed a fair chunk of the report, and I would have a good knowledge base of the issue. If not, I will likely research the ethics of Microsofts racist twitter AI, as delving into internet culture will be fun.