§3.2.1 Specific mitigation measures
27. Specifically, mitigation measures aimed at addressing systemic risks to electoral processes should include measures in the following areas: a) Access to official information on the electoral process. To improve voter turnout and prevent the spread of misinformation, disinformation and FIMI on the electoral process itself, best practice for providers of VLOPs and VLOSEs is to facilitate access to official information concerning the electoral process, including information on how and where to vote, based on official information from the electoral authorities of the Member States concerned. Such information could be provided for example by means of information panels, banners, pop-ups, search interventions, links to websites of the electoral authorities, specific election information tabs or a dedicated part of the platform. When designing and implementing such mitigation measures, the Commission recommends that providers of VLOPs and VLOSEs take principles such as inclusiveness and accessibility into account. b) Media literacy initiatives. Best practice for providers of VLOPs and VLOSEs is to collaborate on, implement, invest and engage in media literacy initiatives and campaigns focussing on elections to foster critical thinking and improve users’ skills in recognizing disinformation and manipulation techniques, also related to generative AI. This could be achieved by: i. Collaborating with local media literacy organisations, as well as relevant associations, groups and networks, financially supporting, sharing and integrating election related initiatives and campaigns on the platform, including by developing joint initiatives. Local media literacy organisations represent a valuable resource in terms of knowledge of local contexts and target audiences. The Commission recommends the use of the network of EDMO and its hubs and the Commission’s Expert Group on Media Literacy to find the relevant organisations at Member State level. ii. Developing and applying inoculation measures that pre-emptively build resilience against possible and expected disinformation narratives and manipulation techniques by informing and preparing users. Such measures should take into account the specific local context where they are carried out and should be complemented by other measures providing reliable information to users. Inoculation measures can take different forms, including e.g., gamified interventions, such as participation in online games on the generation of disinformation which encourages a critical reflection on the tactics used to influence (25), video or other types of content (26), and where feasible, should be implemented in-app to foster ease of access. iii. When designing media literacy campaigns, it is recommended that providers of VLOPs or VLOSEs take into account specific narratives as well as tactics, techniques and procedures (TTPs) that contribute to systemic risks that are likely to occur before, during and after an election, in line with the approach of adapting mitigation measures to the relevant national context and target audiences. c) Measures to provide users with more contextual information on the content and accounts they engage with. Examples include: i. Fact-checking labels on identified disinformation and FIMI content provided by independent fact-checkers and fact-checking teams of independent media organisations. Fact-checking coverage should extend across the EU and its languages, inter alia through strengthening the cooperation with local fact-checkers during election periods, integrating and showcasing election-related fact-checking content as well as employing mechanisms to help increase the impact of these on audiences. Fact-checking labels should be accessible and written in easily understandable language. ii. Prompts and nudges urging users to read content and evaluate its accuracy and source before sharing it. iii. Clear, visible, and non-deceptive indications of official accounts, as well as accounts providing authoritative information on the electoral process, such as the accounts of electoral authorities, including the basis on which such verification is established. The criteria that lead to an “official” label for an account should be made easily available and provided in easily understandable language, to prevent such indications from giving credibility to accounts impersonating official accounts, such as those of electoral authorities. iv. Clear, visible, and non-deceptive labelling of accounts controlled by Member States, third countries and entities controlled or financed by entities controlled by third countries. v. Tools and information to help users assess the trustworthiness of information sources, such as trust marks focused on the integrity of the source based on transparent methodologies and developed by independent third parties. vi. Other tools to assess the provenance, edit history, authenticity, or accuracy of digital content. These help users to check the authenticity or identify the provenance or source of content related to elections. vii. Establish effective internal measures to counter misuse of any of the above procedures and tools, in particular the abuse of the verification process for labelled accounts and content. d) Recommender systems can play a significant role in shaping the information landscape and public opinion, as recognised in recitals 70, 84, 88, and 94, as well as Article 34(2) of Regulation (EU) 2022/2065. To mitigate the risk that such systems may pose in relation to electoral processes, providers of VLOPs and VLOSEs should consider: i. Ensuring that recommender systems are designed and adjusted in a way that gives users meaningful choices and controls over their feeds, with due regard to media diversity and pluralism; ii. Establishing measures to reduce the prominence of disinformation in the context of elections based on clear and transparent methods, e.g. regarding deceptive content that has been fact-checked as false or coming from accounts that have been repeatedly found to spread disinformation; iii. Establishing measures to limit the amplification of deceptive, false or misleading content generated by AI in the context of elections through their recommender systems; iv. Regularly assessing the performance and impact of recommender systems and addressing any emerging risks or issues related to electoral processes, including by updating and refining policies, practices, and algorithms; v. Establishing measures to provide transparency around the design and functioning of recommender systems, in particular in relation to the data and information used in designing systems that foster media pluralism and diversity of content, to facilitate third party scrutiny and research; vi. Engaging with external parties to conduct adversarial testing and red team exercises on these systems to identify potential risks such as risks stemming from biases, susceptibility to manipulation, or amplification of misinformation, disinformation, FIMI or other harmful content. e) Political advertising. Providers of VLOPs and VLOSEs are advised to prepare for the entry into application of Regulation (EU) 2024/900 on the transparency and targeting of political advertising (“Regulation on Political Advertising”) and to take particular care to consider the provision on non-discrimination (Article 5(1)) which will enter into application 20 days after the Regulation’s publication on 20 March 2024 (27). All online platform providers are responsible for ensuring that this provision is complied with. This Regulation defines political advertising as the preparation, placement, promotion, publication, delivery, or dissemination of messages by, for or on behalf of political actors, unless they are of a purely private or a purely commercial nature; or which are liable and designed to influence voting behaviour or the outcome of an election, referendum, or a legislative or regulatory process, at EU, national, regional or local level. Providers of VLOPs and VLOSEs are encouraged to take the definitions provided by the regulation into account, when applying these guidelines. If a provider of a VLOP or VLOSE offers the possibility to place political advertisements on its service, the Commission recommends that, when complying with the obligations laid down in Article 26 of Regulation (EU) 2022/2065 and in line with the forthcoming regulation on the transparency and targeting of political advertising, these are labelled in a clear, salient and unambiguous manner and in real time to allow users to understand that the content displayed contains political advertising. In addition, the labels applied should remain in place when shared by users on the same platform. The Commission also recommends that providers of VLOPs and VLOSEs align their policies to the Regulation on Political Advertising in advance of its entry into application which complement the legal obligations stemming from Articles 26 and 39 of Regulation (EU) 2022/2065, and in particular in the following areas: i. Provide users with information about the political advertisements they see, such as the sponsor identity and, where applicable, the entity ultimately controlling the sponsor; the period during which the political advertisement is intended to be published, delivered or disseminated; the aggregated amounts and the aggregated value of other benefits received by the providers of political advertising services; as well as meaningful information about the main parameters used to determine the recipient to whom the advertisement is presented. ii. Maintain a publicly available, searchable repository of political ads, updated in as close as possible to real-time. This shall include, as a minimum, the total number of recipients of the service reached and, where applicable, the aggregate numbers broken down by Member State for the group or groups of recipients that the advertisement specifically targeted, as mandated by Article 39 of Regulation (EU) 2022/2065, and could also include e.g., the aggregated amounts and the aggregated value of other benefits received by the providers, the number of impressions and the geographical areas in which the ad was presented. iii. When they do not allow political advertising on their services, have efficient verification systems in place and take the necessary actions to ensure that the decision is appropriately enforced. iv. Ensure that there are adequate policies and systems in place to prevent the misuse of advertising systems to disseminate misleading information, disinformation and FIMI with regards to electoral processes, including deceptive generative AI content. f) Influencers can have a significant impact on the electoral choices made by recipients of the service, as they are increasingly involved in facilitating political debate online. In this regard, and in order to increase transparency, providers of VLOPs and VLOSEs should: i. Provide a functionality to allow influencers to declare whether the content they provide is or contains political advertising, including the sponsor identity and, where applicable, the entity ultimately controlling the sponsor; the period during which the political advertisement is intended to be published, delivered or disseminated; the aggregated amounts and the aggregated value of other benefits received by the providers of political advertising services; display period, as well as meaningful information about the main parameters used to determine the recipient to whom the advertisement is presented. ii. Ensure that other recipients of the service can identify in a clear, salient and unambiguous manner and in real time, including through prominent labelling, that the content provided is or contains political advertising, as described in the influencer’s declaration. g) Demonetisation of disinformation content. The Commission recommends that providers of VLOPs and VLOSEs have targeted policies and systems in place to ensure that the placement of advertising does not provide financial incentives for the dissemination of disinformation and FIMI with regards to electoral processes and hateful, (violent) extremist or radicalising content that can influence individuals in their electoral choices. h) Integrity of Services. Providers of VLOPs and VLOSEs should put in place appropriate procedures to ensure the timely and effective detection and disruption of manipulation of the service when this has been identified by them as a relevant systemic risk, taking into account the best available evidence. For example, they may include in their terms and conditions specific rules against the creation of inauthentic accounts or botnets (which may include automated, partially automated, or non-automated accounts), or deceptive use of a service. i. The Commission recommends that providers of VLOPs and VLOSEs develop and enforce their rules, preventing deception through impersonation of candidates, the deployment of deceptive manipulated media, the use of fake engagements, non-transparent paid messages, or non-transparent promotion by influencers, as well as coordination of inauthentic content creation or behaviour. ii. The Commission recommends cooperation between the relevant teams of different providers of VLOPs and VLOSEs to identify common threats and to counter cross-platform disinformation campaigns, FIMI activity or hateful, (violent) extremist or radicalising activity that can influence individuals in their electoral choices and migration of malicious actors (see section 3.4 on cooperation with national authorities, independent experts, and civil society organisations). 28. Considering the evolving nature of the understanding of systemic risks to electoral processes, mitigation measures should be tied to rigorous and critical analysis, testing and review of their intended and potentially unintended impact. As such, effective mitigation measures should be based on the best available information and scientific insights. The Commission recommends that providers of VLOPs and VLOSEs pro-actively design, evaluate, and optimise conceptually valid performance metrics for the effectiveness of mitigation measures, for example via A/B testing of feature and design choices. These performance metrics should be analysed as part of providers’ risk management framework and set to measure the success of relevant mitigation measures during a particular election. These metrics should be SMART (specific, measurable, achievable, relevant and time-bound) and they should be both qualitative and quantitative.
← 3.2 · All articles · 3.2.2 →
Source: EUR-Lex (Cellar) · retrieved 2026-09-07