Electoral Computational Constitutionalism: the limits of algorithmic transparency in the targeting and delivery of political advertising under Regulation (EU) 2024/900

Abstract

The free formation of political will, free from undue interference, constitutes the basic premise of our democratic system. Political advertising, as a form of electoral propaganda, must adhere to the principles of objectivity, transparency, equality, and respect for political pluralism, so that it does not distort the electorate's decision-making process or compromise the authenticity of the vote.

Low-latency multimodal artificial intelligence systems can potentially undermine the traditional guarantees of the integrity, neutrality, and equality of the electoral process, as well as impair the free formation of political will, if their technical capabilities are not limited, contextualized, and parameterized.

In the Spanish legal system, the landmark Constitutional Court Judgment 76/2019 declared the unconstitutionality and nullity of Article 58 bis of the Organic Law of the General Electoral Regime, which authorized political parties to collect personal data regarding citizens' political opinions within the framework of their electoral activities. The Court identified a regulatory insufficiency in the enabling law due to a lack of purpose definition, the absence of predictability regarding the scope of restrictions, and the non-existence of adequate guarantees required for the processing of special categories of data (Art. 9 GDPR), all to the detriment of the essential content of the fundamental right to personal data protection (Art. 18.4 CE) in connection with the principle of legal reservation (Art. 53.1). The Constitutional Court found it unnecessary to rule on the potential violation of ideological freedom (Art. 16 CE).

The full applicability, since October 10, 2025, of Regulation (EU) 2024/900, which chooses to establish, under strict transparency and supervision requirements, the conditions for the use of targeting and ad delivery techniques involving the processing of personal data in the field of political advertising, compels us to question whether the supervision mechanisms provided for in its Articles 14 to 16 (regarding the periodic submission of information by publishers in their management reports, establishing notification mechanisms for potential non-compliant ads to the competent national authorities), as well as the technical transparency specifications established in Implementing Regulation (EU) 2025/1410, are sufficient in the face of the cognitive asynchrony derived from algorithms whose structural opacity places them beyond the reach of the average citizen's critical judgment.

Given this uncertain scenario, and starting from the tension between the opacity of algorithmic targeting and the constitutional requirement of a free vote, this paper analyzes the role that the Spanish Data Protection Agency is called to play, in its capacity as an independent supervisory authority under the GDPR and as the competent authority designated in accordance with Article 22 of Regulation (EU) 2024/900. Furthermore, it examines whether the institutional safeguards provided -a new model of transparency and institutional supervisión- will be legally sufficient to guarantee the voter's cognitive autonomy as a prerequisite of the right to data protection in electoral contexts, or whether they must be complemented with additional measures such as independent algorithmic audits, obligations for material explainability, and mechanisms for individual citizen defense against algorithmic targeting.