Algorithmic Bias and the (False) Promise of Numbers. (19th June 2021)

• Record Type: Journal Article
• Authors: Fejerskov, Adam Moe
• Other Names: Fukuda‐Parr Sakiko guestEditor.; Storeng Katerini T. guestEditor.; Mahajan Manjari guestEditor.; Venkatapuram Sridhar guestEditor.
• Is Part Of: Global policy. Volume 12(2021)Supplement 6
• Page Start: 101
• Page End: 103
• Publication Date: 2021-06-19
• Subjects: Globalization -- Periodicals; International relations -- Periodicals; World politics -- Periodicals; 327.1705
• DOI: 10.1111/1758-5899.12915
• Languages: English
• ISSNs: 1758-5880

Abstract

Abstract: Advances in AI and machine learning systems have given rise to a contemporary euphoria of progress. This commentary discusses the challenges posed by advances in artificial intelligence, or more precisely the increasing usage of algorithmic systems, in global health, sometimes carried forward by an almost blind faith in numbers and automation. Yet we share no common definition of fairness in AI and global health; regulation and regulatory oversight of algorithms seem impossible; and transparency is highly unlikely as algorithms themselves can represent immense commercial value. As we cannot erase the risk of bias in any system that emulates or interacts with our world, the grand challenge becomes to balance innovation with fairness and equality. Abstract : As we cannot erase the risk of bias in any system that emulates or interacts with our world, the grand challenge becomes to balance innovation with fairness and equality.