Author: Arathi Sethumadhavan, Joe Garvin
Publisher: World Economic Forum
Publication Year: 2022
Summary: The following article describes how now that the field of artificial intelligence is growing, data scientists and executives can benefit from looking to the aviation industry, a field known for its emphasis on regulation and safety measures, as a standard for improving data ethics. Like the aviation industry, AI involves high-risk activities due to the large scope and reaching impacts of the models created and applied. Although many of these changes in the aviation industry did not occur until prompted by disasters, the article suggests that the AI industry can learn from the aviation industry’s mistakes and begin implementing reforms now, specifically comprehensive training programs, regulation and standardization, and open sharing of information. Training programs effectively equip individuals with the skills necessary to utilize their resources efficiently and safely, whether this entails pilots and flying planes or data scientists and building models. Often, those creating algorithms may have the technical expertise but lack knowledge about ethical use, so programs need to be comprehensive to promote the protection of those impacted by the models. Similar to the use of shared guidelines in aviation that ensures pilots and technicians maintain industry-wide licenses and uphold a standardized process, AI can benefit from adopting standard policies and requirements to foster consistency and minimize bias. Lastly, AI can benefit from increased transparency about the work being done and greater encouragement of feedback from users. Creating a culture that appreciates and considers input can improve the models being built while building trust amongst users.