DISCUSSION PAPER WORKER MANAGEMENT THROUGH AI From technology development to the impacts on workers

One of the primary reasons why economists tend to focus more on quantity rather than quality is the absence of a measurement approach capable of quantifying several key factors: (i) the degree of technological penetration of such technologies, (ii) the specific scopes of their applications, and (iii) the functions and occupations most susceptible to exposure. [...] These range from privacy infringements to intrusions into the delicate balance between work and personal life, resulting in the potential alteration of individual personalities and the adoption of artificial behaviours to appease the algorithms, all of this resulting in an increase of stress and anxiety in workers (EU-OSHA, 2022b). [...] This measure quantifies the degree of similarity between two text corpora: one derived from the set of identified patents and the other from the descriptions of human functions and tasks performed by workers, using the O*NET dataset. [...] Additionally, the lack of transparency and information asymmetry concerning the use of AI devices can hinder worker agency and potentially heighten the tacit acceptance of these technologies and the related OSH risks. [...] Figure A 1: Structure of the AIWM text query Source: Authors’ elaboration From AIWM patents to task and occupation exposure Following the identification of the target patent pool, we link the textual content of these patents with tasks and human functions found in the O*NET dataset.