Nearly 3 Million Formal Jobs in Mexico Exposed to AI

México, ¿Cómo Vamos? reports that 2.9 million workers, 4.9% of Mexico's labor force, hold administrative and office jobs with medium or high exposure to generative AI, concentrated disproportionately in the formal sector. Affected stakeholders include HR technology providers, corporate employers in finance and administrative services, and policymakers pursuing Mexico's 2030 formalization targets. 

Nearly 2.9 million workers in Mexico, equivalent to 4.9% of the national labor force, hold jobs with medium or high exposure to GenAI, according to a study by México, ¿Cómo Vamos? (MCV). The analysis finds that this exposure concentrates in administrative and office occupations, placing some of the formal labor market's most stable jobs within reach of automation.  

The research, based on first-quarter 2026 microdata from the National Occupation and Employment Survey (ENOE) and the ILO-NASK GenAI exposure index, breaks each occupation into discrete tasks to determine which can be executed or supported with current AI capabilities. Researchers and professionals in mathematics, statistics, and actuarial science, along with information technology coordinators and accountants and auditors, appear among the higher-income roles the study flags as exposed.  

The finding lands as Mexico's formal-informal employment gap remains stagnant. Informality reached 54.8% of the occupied population in the first quarter of 2026, MCV reports, a level barely changed over two decades and above the threshold the study associates with the government's 2030 formalization target. Among occupations with the least AI exposure, by contrast, informality climbs to 66%, underscoring that AI is reaching precisely the segment of the labor market where formal employment, and its associated social security benefits, is most concentrated.

Mexico's overall exposure remains lower than in economies such as the United States, given a labor structure still anchored in physical and in-person work that current AI systems cannot easily replicate. MCV lists 10 occupations with the highest exposure levels: general administrative support staff, data entry clerks, office machine operators, telephone salespeople, accountants and auditors, secretaries, in-person receptionists and information clerks, telephone customer service agents, travel agents, and customs brokerage staff. Tasks such as data entry, appointment scheduling, and document processing rank as highly automatable because of their structured, repetitive, rule-based nature, the report notes, while functions requiring complex contextual judgment, such as financial statement certification, still call for human oversight. Occupations with the lowest exposure include domestic workers, street cleaners, ambulant food vendors, primary school teachers, physicians, and nurses, reflecting the physical and interpersonal demands of these roles.

The findings echo a wider readiness gap already documented in Mexico's talent market. Mexican workers report using AI for an estimated 30% of their tasks today, a figure expected to rise sharply within two years, yet most employees still lack the credentials needed to validate their new skills, a gap that threatens to slow the country's transition toward higher-value work, according to an ETS study . Closing that certification and training gap, alongside targeted formalization policy, will likely determine whether AI adoption in Mexico's office-based occupations displaces formal employment or complements it.  

Taken together, the findings point to a labor market where AI exposure and job formality increasingly overlap, a shift with direct implications for the administrative, financial, and customer service functions that anchor Mexico's formal employment base. MCV researchers frame the challenge less as job loss and more as a call for targeted training and formalization policy, arguing that the pace of skills development, not the technology itself, will determine whether AI adoption strengthens or erodes the formal jobs it is beginning to reach.