US campaigns have spent 137 million dollars on at least 220 ads produced or altered with artificial intelligence ahead of the midterm elections on November 3, 2026, compared with less than 50,000 dollars across the entire 2024 cycle, according to the Wesleyan Media Project report released on October 21.
The figure arrives alongside a change of scale in the mechanism. An El País feature published on October 21, 2026 documents that electoral manipulation no longer depends on isolated pieces and has shifted to autonomous agents, capable of opening accounts, producing videos and holding conversations without human intervention. The text attributes campaigns of this kind to Iran, China and Israel during 2026. The pattern described is not an isolated incident but a low-cost production chain. For Mexico, which renews the Chamber of Deputies in 2027, the operational point is the same: verify the origin of synthetic content before it circulates.
Of the total counted as of October 11, more than 70% of the ads included no disclosure whatsoever about the use of AI, according to the same report. El País describes the arithmetic of the leap: in 2016 the Russian operation on social networks required some 400 Internet Research Agency employees; today, half a dozen people and one computer replicate the scheme. It adds two documented cases: an Iranian campaign that gathered nearly 80,000 followers in the first half of 2026, and the Israeli firm IntelEye, made up of former NSO Group members, which used the DeepSeek model to operate fake accounts on Facebook, Instagram, TikTok and X. Travis Ridout, co-director of the Wesleyan Media Project, notes that this volume makes it harder to vote on the basis of facts and that campaigns spend more time debunking than presenting proposals.
The date is November 3. What follows is measurable: how many of the 220 pieces are withdrawn, how many are debunked, and with what traceability their origin is identified. The precedent leaves the region's electoral processes with a concrete checklist: a count of pieces, the percentage of disclosure and attribution of origin.
This article was written with the assistance of artificial intelligence from verified sources and reviewed by a human editor before publication.