Mexico heads to 2027 with no rules on AI in campaigns
17 governorships, 500 federal seats and over 19,000 posts at stake.

Mexico is already in the electoral process toward June 6, 2027, when voters will choose 17 governorships, all 500 federal deputies (300 by majority and 200 by proportional representation), municipal governments in 30 states, Mexico City’s 16 boroughs and 31 state legislatures, more than 19,000 offices in total. The process began in September without specific rules on the use of artificial intelligence in campaigns.
That is the warning from Karina Vaquera Montoya, a councilor at the State of Mexico’s Electoral Institute, writing in La Crónica de Hoy. The electoral reform proposed by the presidency in March 2026 sought visible labeling of AI-made propaganda, but that provision was not approved. The INE issued guidelines only for newscasts: they recommend telling audiences when content is altered with AI and verifying information before broadcasting it, but they do not reach the rest of the digital environment.
The councilor describes a major legislative gap that could enable false narratives and more disinformation online, and proposes that electoral institutes set up digital monitoring to detect negative narratives. Governorships at stake include Aguascalientes, Baja California, Chihuahua, Guerrero, Michoacán, Nuevo León, Querétaro, Sinaloa, Sonora and Zacatecas, among others.
There is movement in Congress. On September 15, Gabriela Jiménez, Morena’s deputy floor leader in the Chamber of Deputies, called for AI to reach 2027 with clear rules, transparency and accountability, in line with President Claudia Sheinbaum’s call. She proposed working with the INE on criteria for AI-generated or altered propaganda, requiring visible labels on spots, videos, ads and images, especially when a candidate’s voice, image or statements are altered, and banning its use for smear campaigns or disinformation.
The Electoral Tribunal has already ruled on other issues in the process, such as maintaining checks that election observers do not run social programs and counting only ballots that show signs of having been folded. But on AI the referee arrives without a rulebook. The experience of Brazil, which defined what a deepfake is, and of the European Union, which requires labeling of manipulated content, shows the issue reaches the courts sooner or later.
For Mexican consultants, the recommendation is to adopt their own standards now: label every synthetic asset, do not alter the voice or image of real people and document the production process. If regulation arrives mid-campaign, those already compliant will not have to pull material; and if it does not, self-regulation becomes a credibility argument in its own right.
Adopt your own AI-labeling standards now: regulation may arrive mid-campaign.
Summary written by the FOCUS editors based on the original story. Rights to the original content belong to its publisher.
FOCUS courtesy space


