Each of these authors make slightly different choices in the computation of their score, depending on: i) the level of analysis: a whole occupation, the tasks within an occupation or the abilities necessary to pursue the occupation; ii) the source of occupational information: job description databases, worker surveys or job vacancies; iii) the source of technological innovation: patent texts, tech. [...] When analysing the ISCO occupations at the two-digit level in Table A4, it becomes evident that the broad category "41 General and Keyboard Clerks" is the only category present in the top 5 for the two different sources of exposure scores. [...] 6 The correlation coefficient, which corresponds to the correlation between the share of female employment in an occupation and the associated technology exposure measure, is 0.44 for the LM score and 0.37 for the ILO score. [...] While the potential GenAI exposure in the literature is judged both at the levels of tasks (Gmyrek et al, 2023) and abilities (Felten et al, 2023a, 2023b), assessing the actual impact of GenAI always happens at the level of specific tasks. [...] Now, when GenAI (like OpenAI’s Advanced Data Analysis) automates some early steps in the data cleaning, exploration and modelling phases, the loss of information could worsen the data scientist’s ability to interpret the results and communicate recommendations in the later steps of the process.
- Pages
- 33
- Published in
- Belgium
