WP121 Web · Page 39

36
  ILO Working Paper 121
– the middle class is the group whose jobs and earnings have the highest levels of overall expo­
sure to GenAI, with many possible directions that this transformation can take.
These findings suggest an important role for government interventions, aimed at minimizing 
disruptions resulting from sudden job losses through job protection measures, and maximizing 
the productive benefits of the transition, for example through equipping workers with founda­
tional skills that can help them keep up with the changing character of jobs and drive the pro­
ductive character of such changes, rather than see their skills become obsolete. The fact that 
some vulnerable groups, such as women and youth, face greater exposure to automation high­
lights the importance of life-long learning so that that workers have the skills to adapt to chang­
es in the world of work. In the short-term, as shown in numerous ILO studies, social protection 
systems can play an important role as macroeconomic shock stabilizers, and reduce the impact 
of transitions for the affected workers and their households at the microeconomic level (ILO, 
2023), especially when their use is combined with skills development programs (ILO, 2023e). In 
the medium and long-term, reducing gender gaps in the exposure to automation would require 
addressing factors that perpetuate occupational segregation by gender, such as gender-based 
social norms (Carranza et al., 2023).
At the same time, our findings show that the shares of jobs that could benefit from a productive 
transformation with GenAI are consistently higher than those with automation risks across all 
LAC countries, ranging between 8 and 12 percent of employment across countries. This is par­
ticularly the case for the jobs in education, health and personal services. In addition, the sectors 
oriented towards customer service (retail, trade, hotels, restaurants, etc.) face an elevated expo­
sure to "the big unknown", which means that a productive augmentation could also be sought 
in these jobs with the right policies and incentives in place. Therefore, a tempting narrative that 
can be constructed based on these statistics is that, in the big picture, more can be gained than 
lost as a net job and economic effect of the transformation.
This is where our analysis provides new information to assess whether the lack of digital infra­
structure could be a buffer or a bottleneck to reap the economic benefits of GenAI. On the one 
hand, our findings show that most workers exposed to GenAI automation are using digital tech­
nologies, which suggests that the potential negative effects may not take long to materialize. On 
the other hand, we find that inadequate digital infrastructure is a major bottleneck to realizing 
the positive effects of augmentation, thereby impacting a significant segment of the labour force 
in the LAC region. Nearly half of the occupations that could potentially benefit from augmen­
tation are hampered by digital shortcomings that will prevent them from realizing that poten­
tial. Specifically, 6.24 percent of jobs held by women and 6.22 percent of those held by men are 
affected due to these gaps. Similar limitations apply to the jobs in the “big unknown” category: 
even though some of them could potentially pivot towards augmentation through increasing 
complementarity between GenAI and the worker in these occupations, the digital gaps will pre­
vent large shares of these jobs fro