Author(s) :

Shanmukh Sriram. Chintapalli, Dr. Madhura Shivaram, PhD.

Volume/Issue :

Abstract :

Since ChatGPT launched in November 2022, artificial intelligence (AI) has rapidly entered the workplace, raising questions about whether its effects are felt equally across different groups of workers. This paper analyzes observed 2022–2025 employment associations alongside BLS 2024–2034 occupational projections to characterize both current patterns and anticipated trends, associations between occupational AI task exposure and employment patterns across demographic groups ‒ and which groups face the greatest risk based on projected trends. Using U.S. government employment survey data from 2022 to 2025, jobs are classified into three categories: those with higher measured AI task exposure where substitution risk may be elevated, those where AI tools appear associated with productivity gains, and those showing minimal measured AI task exposure. A machine learning model identifies which factors best predict whether a job category will grow or shrink. Overall employment remained stable, but this masks meaningful differences: women make up around 70% of workers in roles with the highest AI task exposure scores, concentrated in office and administrative work, where substitution risk may be elevated; workers without a college degree show nearly zero projected job growth, while those with advanced degrees are projected to gain over 9%. Education level and occupation type are closely associated with changes in employment. This paper concludes that employment patterns associated with AI exposure do not appear equal across all workers ‒ observed patterns differ by occupation and education. Because the study covers a short time window, some trends may reflect post-pandemic recovery rather than AI alone, and future research should apply stronger causal methods over longer timeframes.