Putting the Pieces Together: Why Linguistic, Media & AI Dynamics, and EU Policy Need to Come Together to Debunk Disinformation

In recent years, disinformation has gone from the occasional shots of “fake news” to a systemic threat that uses linguistic manipulation, media mechanisms, artificial intelligence, and transnational political strategies to form public opinion and undermine democratic processes. Two strands of research, one dealing with the “micro” mechanisms of disinformation and the other with high-level policy and governance responses, have often remained separate. It is time to bring them together.

The Anatomy of the Threat: Language, Media Mechanisms, and generative AI

Disinformation campaigns are achieved through various tiers, which are interrelated, making it very effective (D’Andrea et al., 2025a). First, at the level of language, manipulative content is used, which is based on emotions and polarization, affecting perceptions rather than informing. It is then spread through the various mechanisms of the media and the platform, which are based on the cognitive biases of the users, who are drawn to sensational and polarizing content. Furthermore, the appearance of generative AI makes this problem particularly important, as it is possible to produce a great deal of very believable false content, which is very difficult to detect (Shukla & Tripathi, 2024; Pilati & Venturini, 2025; Shoaib et al., 2023). This, of course, has very important consequences for the public sphere, affecting political opinions and decision-making processes.

The Role of AI: As Risk and as Tool 

AI is a double-edged sword. On the one hand, generative AI facilitates large-scale disinformation campaigns (Shoaib et al., 2023; Feuerriegel et al., 2023). On the other hand, AI tools can detect, track, and counter manipulative content if used in a responsible manner (Curtis et al., 2025; Truică & Apostol, 2022; Leite et al., 2025). Techniques include cross-lingual detection, network analysis, or automated fact-checking pipelines (Peña-Alonso et al., 2025; Shukla & Tripathi, 2024). Moreover, explainable AI systems can contribute to the detection of deepfakes and the analysis of network diffusion patterns (Curtis et al., 2025; Truică & Apostol, 2022).

Why Policy Alone is not Enough — But Policy is Still Essential

In terms of policy, the European Union has put in place programs like the European Digital Media Observatory (EDMO) and the Digital Services Act to increase media literacy, accountability, and transparency among its member states (European Commission, 2024; Sánchez Gonzales et al., 2024; D’Andrea et al., 2025b). Collaborative networks that include fact checkers, journalists, and academic institutions are integral in counterbalancing this at scale (Frau‑Meigs et al., 2025). They are usually initiatives that respond to the problem and, thus, control the consequences of disinformation without addressing its cognitive, linguistic, or algorithmic roots (Leite et al., 2025). The policies that are formed without being backed by research in domains such as emotional framing, persuasive strategies, and AI-driven amplification have the risk of being too broad or ‍‌‍‍‌scattered.

Merging Micro and Macro-Toward a Holistic Strategy

Effective mitigation of disinformation must be addressed both at the micro and at the macro levels:

• Micro-level vigilance: citizens must become capable of spotting manipulative narratives, emotionally framed and/or algorithmically amplified content (Arribas et al., 2025).
• Macro-level governance: this would necessitate regulatory frameworks that provide for transparency, accountability of platforms, access to data to allow for independent analysis, and the overall coordination between EU member states and fact-checking networks (Pilati & Venturini, 2025; European Commission, 2024).