The topic of the second episode of the second season of the „Efekt Sieci” podcast is algorithms and systems based on automated decision-making. In the US, the problem with racist or sexist biased algorithms is already starting to take a real toll, writes the Financial Times. It is for these reasons that systems in Idaho have decided to reduce the amount of Medic-aid benefits paid to 4,000 disabled people. Amazon also faced this problem when deciding in which regions to roll out its new Prime service, which entitles consumers to faster delivery and special promotional offers. Amazon’s algorithms decided in advance that implementing the service made no sense in neighbourhoods inhabited mainly by ethnic minorities, such as Roxbury in Boston or the South Side in Chicago. At the same time, they stated that Prime should go to the so-called white districts. Is there anything to fear?
The expert for the episode is Dr. Joanna Mazur, analyst at DELab UW, Faculty of Law and Administration, UW, who answers questions from Dr. Justyna Pokojska about:
- the implications of systems based on automated decision-making processes,
- the need to come to terms with the distortions and side-effects inherent in the operation of automatic decision-making processes,
- places and industries where machines are already replacing humans in decision-making,
- the use of algorithms in the public sector,
- scenarios in which automated decisions become part of the law,
- security and impartiality of decision-making in this way,
- control of the software and its developers – ethics, transparency and accountability of decisions taken with their help.