Argentina 2023 General Election

The counts were 1,675 tables for Juntos por el Cambio, 1,669 for La Libertad Avanza and 1,652 for Union por la Patria, out of 7,061 tables where at least one force showed zero. The three counts tracked one another district by district: in Buenos Aires province the figures were 695 tables for Bullrich, 702 for Milei and 708 for Massa; in Salta 186, 180 and 174; in the City of Buenos Aires 112, 112 and 113. National Electoral Director Marcos Schiavi said 1,544 tables could not be tallied at all in the provisional count because their telegramas did not arrive in time, which produces a zero for every candidate on those tables.

Disposition 4/2023 of the National Electoral Directorate funded each qualifying grouping to print its own ballots at 2.5 per registered voter per district, with a reference unit value of 2.92 pesos. Accredited observers reported that the contingency ballots held by polling-station authorities were needed where a party's ballots ran out, attributing the shortages either to the volume of voters or to ballots being removed from the voting booth. Congress replaced the system with a state-printed single paper ballot in Law 27,781 of 2024.

IDEMOE deployed 30 observers and experts across six districts - the City of Buenos Aires and the provinces of Buenos Aires, Cordoba, Mendoza, Santa Fe and Tierra del Fuego - plus digital monitoring in five more. It reported that over 82% of polling stations opened between 8:00 and 8:30 and that 98% of establishments handled the count well to very well, but recorded that its accompaniers had to leave before the 18:00 close and could not observe the closing, the preparation of the telegramas or the tally. IDEMOE described this as a step back from previous elections and as a limit on how comprehensive an integrity assessment it could make.

Freedom House recorded an unprecedented degree of online content manipulation around this election, including deepfakes, produced by both leading candidates and their supporters, and stated that after the first round Milei and supporters of his party used digital platforms to spread baseless allegations of electoral fraud. Argentine reporting documents the same fraud claims circulating from images of provisional-count telegramas.

The court said that every complaint must identify the person making it and that an anonymous one is not valid, and that channels set up by parties or candidates which do not verify identity have no legal effect and can obstruct the investigation of a genuine irregularity. The statement was reported as a response to a website La Libertad Avanza had opened inviting anonymous reports of irregularities; the court's wording is general and does not name a party. Complaints of election offences run through the electoral justice system under the National Electoral Code.

The National Electoral Chamber said the zero-vote telegramas were preparation errors of a kind always present in a small proportion and corrected in the definitive count, and comparable counts exist in earlier elections: 1,624 and 1,546 zero-vote tables for the two leading candidates in 2015, and 212 and 197 in 2019. The National Electoral Directorate reported that only 125 telegramas, 0.12% of the presidential total, showed zero votes for La Libertad Avanza alone. As of 1 November 2023, the day before the first-round definitive count concluded, no appeal questioning the validity of any polling station's count had reached the National Electoral Chamber from any party.

The provisional count published by the National Electoral Directorate gave Union por la Patria 36.69% and La Libertad Avanza 29.99% in the first round; the definitive count gave 36.78% and 29.99%. In the run-off the provisional count gave 55.69% to 44.31% and the definitive count 55.65% to 44.35%, a shift of four hundredths of a point on a margin of about 2.96 million votes.

The provisional count for both rounds was run by Indra under a contract awarded by the National Electoral Directorate in April 2023, after a tender whose draft specification was published a month before the deadline and drew 75 comments from parties, technology firms and civil society; Indra bid 6,667,412,627 pesos against one other technically qualified bidder. The complaints that followed concerned telegramas that were not transmitted in time rather than transmitted totals that were altered, and the judicial count resolved those tables. No cited source identifies an error introduced by the provisional-count system itself.

Freedom House attributed the manipulated and AI-generated content it recorded to the candidates and their supporters and reported no foreign state involvement. Neither the electoral authorities' statements nor the domestic observation report cited here identifies a foreign actor targeting the 2023 Argentine election.

Freedom House documented the scale and character of the manipulated content but did not measure any effect on either round, and no Argentine authority assessed its impact on the result. A finding about online content establishes what circulated, not that it changed how people voted or that it touched the count.