2 ms·
How Algorithmic Experiments Harm People Living in Poverty
- downrightmike 9y agoIn regards to governmental programs like Unemployement Insurance "The programs more than paid for themselves. The reduction in UI benefits paid plus the increase in tax receipts from faster reemployment were more than enough to pay for the cost of providing the search assistance." "The statisitcal targeting of these search-assistance programs was crucial to their cost savings. The costs of government support programs can balloon if they are opened to too large a class of individuals" There have been provisions as far back as 1989 in the US federal government that mandate the use of control groups and randomized trials. Over the past nearly 30 years these tests have lead to better service to more people than would be available without the algorithms. [1] SuperCrunchers by Ian Ayers https://books.google.com/books?id=dJXZ-EzX6gEC&pg=PA67&lpg=PA67&dq=ian+ayers+government+programs+using+statistics&source=bl&ots=hDoA-CFmFX&sig=SBOKi_boMN4TZRPii18ukIYWLi8&hl=en&sa=X&ved=0ahUKEwin2L2N2_PYAhUFv1MKHX2xB4sQ6AEISTAF#v=onepage&q=ian%20ayers%20government%20programs%20using%20statistics&f=false https://books.google.com/books?id=dJXZ-EzX6gEC&pg=PA67&lpg=P...
- YeGoblynQueenne 9y agoAhem. "If you want a picture of the future, imagine a statistical model rejecting an online application, for ever".