TY - JOUR
T1 - Large language models impact on agricultural workforce dynamics
T2 - Opportunity or risk?
AU - Marinoudi, Vasso
AU - Benos, Lefteris
AU - Villa, Carolina Camacho
AU - Kateris, Dimitrios
AU - Berruto, Remigio
AU - Pearson, Simon
AU - Sørensen, Claus Grøn
AU - Bochtis, Dionysis
N1 - Publisher Copyright:
© 2024
PY - 2024/12
Y1 - 2024/12
N2 - Motivated by the rapid advancement of large language models (LLMs), this study explores the potential impact of them on agricultural labor market. Starting from the task level of each of the 15 selected occupations, their exposure to LLMs was assessed by rating the extent to which the required abilities are aligned with those of LLMs, taking also into account the importance of the abilities in each occupation. Findings indicate that while LLMs can significantly enhance cognitive functions, they cannot fully replace the physical, psychomotor, and sensory abilities. As a consequence, while certain tasks are either partially or highly susceptible to LLM implementation, a considerable proportion, involving manual responsibilities, remains largely unaffected. It was seen that occupations heavily reliant on data are at greater risk of substitution. Conversely, some occupations will probably experience an augmenting effect, as LLMs will automate certain cognitive routine tasks, freeing up human workers to focus on more creative non-routine aspects. Furthermore, a negative correlation between exposure to LLMs and exposure to robotization was found highlighting the interconnected dynamics between these two variables within the analyzed context. In conclusion, although LLMs can offer substantial benefits, their integration necessitates careful consideration of their inherent limitations to maximize efficacy and mitigate risks in the agricultural sector.
AB - Motivated by the rapid advancement of large language models (LLMs), this study explores the potential impact of them on agricultural labor market. Starting from the task level of each of the 15 selected occupations, their exposure to LLMs was assessed by rating the extent to which the required abilities are aligned with those of LLMs, taking also into account the importance of the abilities in each occupation. Findings indicate that while LLMs can significantly enhance cognitive functions, they cannot fully replace the physical, psychomotor, and sensory abilities. As a consequence, while certain tasks are either partially or highly susceptible to LLM implementation, a considerable proportion, involving manual responsibilities, remains largely unaffected. It was seen that occupations heavily reliant on data are at greater risk of substitution. Conversely, some occupations will probably experience an augmenting effect, as LLMs will automate certain cognitive routine tasks, freeing up human workers to focus on more creative non-routine aspects. Furthermore, a negative correlation between exposure to LLMs and exposure to robotization was found highlighting the interconnected dynamics between these two variables within the analyzed context. In conclusion, although LLMs can offer substantial benefits, their integration necessitates careful consideration of their inherent limitations to maximize efficacy and mitigate risks in the agricultural sector.
KW - Abilities
KW - Artificial intelligence
KW - Complementary
KW - Human-machine Interaction
KW - ONET
KW - Skills, Substitution
UR - https://www.scopus.com/pages/publications/85211026775
U2 - 10.1016/j.atech.2024.100677
DO - 10.1016/j.atech.2024.100677
M3 - Journal article
AN - SCOPUS:85211026775
SN - 2772-3755
VL - 9
JO - Smart Agricultural Technology
JF - Smart Agricultural Technology
M1 - 100677
ER -