Artificial intelligence impacts on organizations and work: a Delphi study with brazilian experts

• Ano de defesa: 2019
• Autor(a) principal: Pauli, Sergi
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Artificial intelligence; Delphi method; Impactos; Impacts; Inteligência artificial; Método Delphi; Organizações; Organizations; Trabalho; Work
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/12/12139/tde-22082019-153846/

We are living a new and emerging technological wave that is mainly based on Artificial Intelligence. It is being led by the great IT corporations and could potentially bring transformation and disruption in large scale to the economy, industries, businesses, organizations, and people in the years to come. Frey & Osborne (2017)\'s research was an important milestone in evaluating the impact of AI and automatization in the future of employment and their key conclusion that 47% of total U.S. employment was at risk of being potentially extinguished in a decade or two had an enormous impact on the mass media. Therefore, the objective of this study is to evaluate what could be the key impacts of Artificial Intelligence on Organizations and Work. In doing so, we scrutinize the authors\' research and propose an alternate ranking of occupation\'s susceptibility based on a different method and grounded on experts\' opinions. We also evaluate Frey & Osborne (2017)\'s key finding regarding to employment impact by technologies and identify key positive and negative qualitative impacts of AI on organizations and work, occupations and labor market. Taking into account the nature of this research, which is forward-looking, experimental and propositional, focused on current and future implications of Artificial Intelligence, we performed field research with experts supported by a Delphi Method, which is complemented by other techniques. Delphi is a robust and proven method commonly used in future research to assess the direction of long-range trends, with special emphasis on science and technology, and their probable effects on our society and our world. Among our key conclusions, we evaluate bottlenecks applied in to the occupation context and compare them to those identified by Frey & Osborne (2017). We also create our susceptibility ranking that takes into account an integration complexity factor, derived from Metcalfe\'s Law, which shows that occupations with less integration complexity, like clerks and assistant positions, are more likely of being replaced, while the ones that demand higher integration of abilities are practically not at risk. These results help in elucidating the current and future situation of this theme and allow us to suggest some possible suppositions. One of the most important is that no occupation will reach the 100% susceptibility index in twenty years, contrary to Frey & Osborne (2017)\'s research, which means that few occupations can be entirely replaced with acceptable quality by machines that combine Artificial Intelligence, Robotics, and related technologies. Yet, our most relevant finding is this research is related to complexity and integration of occupations. Technologies may emulate individual abilities to a higher extent in the future, but more important than that is being able to harmonically combine these capabilities and make them work together with synergy to achieve even basic tasks of occupations. This integration challenge in association with Autor (2015)\'s Polanyi\'s paradox corroborates the fact that no matter how advanced technology might be in a specific ability, it takes more than that for machines to successfully replace humans in an occupation, which we understand confirms the future scenario of collaboration, complementation and synergy between humans and machines, rather than the replacement and displacement