ALGORITHMIC GOVERNANCE AND ELECTORAL MANIPULATION: AI SYSTEMS IN ROMANIAN POLITICAL CAMPAIGNS AND DEMOCRATIC VULNERABILITIES

Abstract

Artificial intelligence and algorithmic recommendation systems have become central instruments in contemporary electoral campaigns (Helbing et al., 2019), fundamentally transforming how political messages are created, targeted, and amplified. Romania, as an EU member state, faces a confluence of threats: technological and legal vulnerabilities unaddressed by 2015 electoral legislation, limited institutional capacity, and escalating sophistication in AI-assisted disinformation. This article analyzes the mechanisms through which algorithms influence electoral behavior, identifies critical gaps in Romanian legal frameworks, and proposes legislative and institutional reforms to protect electoral integrity in the AI era. Examining the 2020 and 2024 presidential elections, this study identifies specific vulnerabilities within the Romanian electorate and comparative regulatory models from other European democracies. The analysis reveals that current constitutional protections for free voting (Article 36) and freedom of expression (Article 30) require urgent judicial interpretation in the context of algorithmic manipulation. The article proposes a comprehensive reform strategy including amendments to the Electoral Code (Law 207/2015), creation of a unified Electoral AI Authority (AAEE), and alignment with EU regulatory frameworks including the Digital Services Act (2022/2065) and GDPR. The research demonstrates that regulatory approaches adopted in Germany, France, and Austria offer viable models for protecting electoral integrity without restricting legitimate political expression or forcing platform withdrawal from Romanian markets.