The Ubiquity and Risk of Algorithms
Mary W. Gray
American University
Algorithms are determining who is arrested and the sentencing of those convicted, as well as decisions on medical care, school admission, insurance, employment, housing, voting,, and other essentials of life today. If the algorithm is patented or classified as proprietary, there may be little recourse by the subject or her/his attorneys should impermissible factors be used in constructing the decision-determining algorithms. This is becoming of increasing concern today because of the wide-spread use of machine-learning to construct complex, multiple versions of algorithms tailored to produce results that may cause physical or mental harm to individuals throughout the world as well as constituting substantial invasions of privacy. Examples from human rights, national and international law, economic development, journalism and government actions will be discussed.