Our results show a fragmented landscape of data transparency across regions and platforms. Large platforms such as Meta, TikTok and LinkedIn concentrate meaningful transparency in jurisdictions with stronger oversight. The disparity is even more pronounced in advertising data, where the regional access gap is particularly stark. Differences also appear across platforms: while Bluesky and YouTube provide relatively strong transparency resources, demonstrating that meaningful access is technically feasible, many platforms offer little or no access even where regulatory requirements exist.
Social media transparency remains poor across both UGC and advertising. Our findings show that widely used platforms such as Discord and Kwai (Kuaishou) offer no data transparency mechanisms for either content type in any of the regions assessed. Meanwhile, major platforms, including those owned by Meta, and LinkedIn, offer only Negligible or Minimal levels of UGC data transparency across all regions, at best. Other platforms, such as Pinterest and Snapchat, provide only Negligible advertising transparency in a limited number of regions. Across the 15 platforms assessed in all regions, ten were rated as at best Deficient in terms of transparency resources for public UGC data, while at least seven platforms received similarly low ratings for advertising data transparency. This means that their transparency resources are either non-existent or subject to severe restrictions that hinder their effective use, such as limited API scope, significant gaps in data coverage, and missing data or filtering capabilities.
Platforms that lead in public UGC transparency offer little or no advertising transparency, and vice versa. Bluesky, which does not have an advertising-based business model, and YouTube are the only platforms classified as Meaningful for UGC data transparency, but YouTube provides only Deficient advertising transparency on average. X, Reddit, and Telegram also present a similar pattern of asymmetry between the two dimensions, favouring UGC data. Conversely, despite their low UGC data transparency scores, Meta’s platforms (Facebook, Instagram, Threads, and WhatsApp) lead in advertising transparency in regulated markets.
Regulation appears to drive social media data transparency, especially in advertising data, but does not guarantee it. Under the DSA, the EU exhibits greater transparency yet still struggles to ensure compliance. X and Snapchat both fall under DSA requirements in the EU, yet provided only Negligible advertising transparency tools at the time of assessment. Although such tools formally exist, they are non-functional and fail to produce results, amounting to merely performative transparency.
The companies’ compliance-driven approach toward transparency results in regional asymmetries. Brazil consistently records the lowest overall transparency scores for both public UGC and advertising data, while the EU achieves the highest scores for advertising transparency. The UK often matches or even exceeds EU levels, which may reflect regulatory spillover effects, whereby functionalities such as Meta Ad Library’s features initially available only in the EU, as well as the availability of TikTok’s Research API and Commercial Content Library, have been extended to the UK. While Meta’s platforms, Google (YouTube), TikTok, and LinkedIn provide Limited or Meaningful transparency tools in the EU, they rate as only Deficient, Negligible, or Not Available in Brazil. This selective transparency, largely concentrated in jurisdictions with stronger regulatory oversight, produces significant disparities in the production of data-driven evidence on societal issues online.
Research capacity on social media is currently severely constrained across both public UGC and advertising data, making it nearly impossible to construct a comprehensive view of the contemporary information ecosystem. The ability to trace how content flows across platforms, how advertising targets audiences, and how information operations exploit the interplay between organic and paid content remains out of reach. Current fragmented data access is not a technical limitation but a pattern of selective disclosure shaped by corporate strategy and jurisdictional compliance. The stark divide between a small number of top performers (Bluesky and YouTube for UGC data, and Meta’s platforms for advertising data in regulated markets) and the vast majority of services shows that data transparency is ultimately a matter of corporate choice, rather than a technical, commercial, or legal constraint.
Transparency cannot be achieved through mandates alone without accompanying standardisation of data access protocols and enforcement measures. Enforcement is only as strong as the data available to verify compliance, and where transparency tools are broken or absent, regulators lack the basic means to audit platform practices. Even under the DSA, platforms such as X, Snapchat, and Pinterest provide only Negligible transparency tools, with X having been fined by the European Commission for non-compliance in 2025 (Marsh and Seiling 2026). The absence of effective penalties or enforcement mechanisms in many jurisdictions further compounds the problem, creating an environment in which non-compliance carries little cost.
The lack of common data access standards means that even well-intentioned regulatory frameworks produce fragmented outcomes. Each platform implements its obligations differently, forcing researchers to navigate a patchwork of incompatible systems, each with its own limitations and restrictions. This lack of interoperability not only raises the cost and complexity of research but also creates systematic advantages for well-resourced institutions in jurisdictions where platforms offer better data access, deepening existing inequalities in global research capacity. For researchers in the Global Majority, who often face the most acute information integrity challenges, these barriers are especially consequential.
Not available (0)
Negligible (1–20)
Minimal (21–40)
Deficient (41–60)
Limited (61–80)
Meaningful (81–100)
| Platform | Brazil | EU | UK |
|---|---|---|---|
 YouTubeVLOP |
|||
 Bluesky |
|||
 X/TwitterVLOP |
|||
 Telegram |
|||
 Reddit |
|||
 TikTokVLOP |
|||
 LinkedInVLOP |
|||
 FacebookVLOP |
|||
 InstagramVLOP |
|||
 Discord |
|||
 Kwai |
|||
 PinterestVLOP |
|||
 SnapchatVLOP |
|||
 Threads |
|||
 WhatsApp |
VLOP Very Large Online Platform designated under the EU Digital Services Act (DSA), subject to enhanced transparency and accountability obligations.
Despite regulatory progress, access to public UGC data remains limited in all regions. Ten of the 15 platforms assessed either do not have any interface or programmatic means for data access, or only provide access at a level we classified as, at best, Deficient across all assessed regions. Discord, Kwai, Pinterest, Snapchat, Threads, and WhatsApp lack any data transparency resource and have therefore been deemed as Not Available. Similarly, transparency on Facebook and Instagram has been classified as Negligible across all regions, as extracting structured data from the Meta Content Library (MCL) API to a users’ devices violates the platforms’ terms and conditions. TikTok only provides a level of Deficient access in the EU and the UK, while LinkedIn offers only Minimal access in the EU.
Apart from these few exceptions, the EU’s greater regulated environment does not appear to provide a substantial advantage in public UGC data transparency compared to the UK and Brazil. In practice, our request for access to LinkedIn’s Research API under Article 40 of the DSA was not approved, and we were only able to rely on its public documentation to assess a limited set of data quality criteria. Similarly, our request for access to X’s API under Article 40 of the DSA was denied. Pinterest and Snapchat also do not provide any programmatic access under Article 40 of the DSA. Instead, they offer curated datasets to selected researchers under their own conditions, which do not comply with the standards outlined in the Materials & Methods section.
Meta has drastically reduced access to public UGC data across its platforms. CrowdTangle, its previous solution for UGC data transparency, offered relatively flexible access but was discontinued in 2024. It was replaced by the MCL, a far more restricted and controlled environment comprising an API and a GUI, to which our access was approved and which we were able to test for this study. However, the MCL API was classified as Negligible, as its terms and conditions prohibit the extraction of structured data to users’ own devices. Through the MCL API, pre-approved users can access and analyse data from Instagram and Facebook only, and solely within a controlled-access environment. By contrast, the MCL GUI allows the extraction of structured raw data from Facebook and Instagram, albeit with significant limitations on the availability of public data. Threads data can be accessed through this interface but cannot be extracted in any form. For WhatsApp, no public UGC data (e.g. from public channels) is accessible via either the MCL API or its GUI.
Ranked in an intermediate position, Reddit and X offer reasonably developed data transparency resources, although both present critical limitations that undermine research efforts. Both platforms were classified as having Limited transparency for UGC, meaning they provide structured access to data but with constraints on volume and/or data quality. As with Meta’s shift from CrowdTangle to the MCL, the transition from Twitter to X marked a significant regression in data transparency in the early 2020s. Once widely regarded as a leading example of openness in social media data access—offering a free, well-documented API extensively used for research—X introduced paid access to its API in 2023 (Mehta 2024). At the time of our evaluation in 2025, monthly subscription costs started at USD 200 and could reach up to USD 5,000 for more extensive research use.
Our request to access X’s API under the provisions of the DSA in the EU was initially submitted in October 2025 and, after multiple follow-up attempts, was definitively rejected in December of the same year, without further justification. As a result, the API could not be effectively used or assessed under these conditions. This experience is consistent with that reported by other researchers and contributes to a broader pattern that ultimately led to the imposition of a fine against X in the EU due to its inadequate transparency practices (Marsh and Seiling 2026).
Among the messaging platforms evaluated, Telegram is the only one that offers any data transparency mechanisms for public UGC. Through its API, Telegram provides free programmatic access to messages in public channels and groups, scoring highly across our criteria for accessibility, completeness, consistency, and timeliness. The main limitations of data access on Telegram are the absence of platform-wide content discovery—requiring prior knowledge of specific channels—and shortcomings in documentation quality, which place it in the Deficient band of our analysis. Discord and WhatsApp offer no mechanisms for researchers to access or extract public UGC data.
In the UK and the EU, TikTok performs moderately in our analysis thanks to its Research API, which is available to researchers in the U.S. and Europe, but not in Brazil. TikTok’s free API access, robust documentation, and data filtering capabilities place it in the Deficient band under our methodology. However, its application and approval process, combined with delays in updating the data made available to researchers, create a misalignment between what is observable on the platform and what is accessible for critical inquiry and research, ultimately limiting its overall performance. Although TikTok is not subject to regulatory obligations to provide this API to UK-based researchers, it has done so, which we hypothesise to be a form of regulatory spillover—part of the “Brussels Effect”, whereby regulatory standards adopted in the EU shape platform practices beyond its jurisdiction (European Parliament 2025; Husovec and Urban 2024).
Bluesky and YouTube emerge as the top performers in the transparency assessment for public UGC data, with both offering free, full-scope programmatic access to such data. Bluesky’s open-source architecture is designed to enable transparency and unrestricted data access, although its documentation received a weaker evaluation. YouTube, in turn, offers a well-structured API with extensive documentation; however, strict quota limits significantly limit access in practice, undermining transparency for large-scale research. Overall, these are the only two platforms classified in the Meaningful band in our analysis.
Even on platforms that provide some degree of transparency, critical bottlenecks continue to constrain social media research. Essential functionalities such as native geographic and language-based filtering are often absent from APIs, while access to historical content remains negligible or not available. As a result, researchers conducting localised analyses must extract substantially larger datasets than necessary and rely on computationally intensive post-processing. These approaches can introduce additional sources of bias and error, while also requiring significant infrastructure—constraints that systematically disadvantage under-resourced researchers, often those best positioned to investigate information integrity challenges in underserved regions and languages. At the same time, restricted access to historical data precludes longitudinal analysis, limiting the ability to examine how phenomena such as disinformation campaigns or hate speech evolve over time. These limitations are not marginal inconveniences, but fundamental constraints on the kinds of knowledge that can be produced about the digital information environment.
| Platform | Brazil | EU | UK |
|---|---|---|---|
| FacebookVLOP |
|||
| InstagramVLOP |
|||
| Threads |
|||
| WhatsApp |
|||
| LinkedInVLOP |
|||
| TikTokVLOP |
|||
| YouTubeVLOP |
|||
| PinterestVLOP |
|||
| SnapchatVLOP |
|||
| X/TwitterVLOP |
|||
| Discord |
|||
| Kwai |
|||
| Reddit |
|||
| Telegram |
VLOP Very Large Online Platform designated under the EU Digital Services Act (DSA), subject to enhanced transparency and accountability obligations.
Bluesky is excluded from the Ads assessment as the platform does not have advertising.
Half of the platforms assessed for advertising transparency were classified as Not Available or Negligible across all regions. Overall, seven of the 14 platforms with an advertising business model fail to meet basic transparency standards—Bluesky was not evaluated as it does not run ads. Discord, Kwai, Reddit, Telegram, and Pinterest provide no mechanisms for structured access to advertising data. X and Snapchat maintain ad transparency tools in the EU under DSA obligations, but these were non-functional at the time of assessment, placing both in the Negligible band.
Advertising data transparency reflects a landscape shaped by both regulatory pressure and voluntary decisions. Unlike public UGC data—where platforms such as Bluesky, YouTube, X, Telegram, and Reddit provide Meaningful or Limited transparency for global access—advertising transparency is marked by pronounced regional disparities. Among platforms that offer some level of ad transparency, the gap between Brazil and the UK and the EU is both stark and consistent. Meta’s platforms achieve a Meaningful classification in the UK but drop to Deficient in Brazil; YouTube declines from Limited in the EU to Negligible in Brazil; and TikTok falls from Limited in the EU to Not Available in Brazil. These patterns indicate that the technical infrastructure exists but is deployed selectively.
Despite the positive effects of regulation, some companies subject to the DSA lack meaningful transparency and were classified as providing Negligible resources in the EU, highlighting the gap between nominal compliance and functional transparency. Pinterest, X, and Snapchat exemplify this pattern. In the EU, our application to X’s API access to advertising data remained pending without response after three months and could therefore not be evaluated, while its GUI export feature returned no data across all tested queries, which are limited to advertiser selection. In Brazil and the UK, X was classified as Not Available, with no transparency tools. Similarly, Snapchat’s EU ad transparency tools proved non-functional: the API consistently returned errors, and the GUI yielded no results regardless of the search queries applied. Pinterest, in turn, offers a GUI-based ad library in the EU and Brazil, but not in the UK, allowing users to browse ads and view basic advertiser and temporal information. However, Pinterest provides no API or structured data extraction capabilities and does not support keyword search, comprehensive filtering, or systematic analysis. A previously available API for advertising data extraction was deprecated in 2022 and has not been replaced.
The highest levels of advertising transparency were observed on Meta’s platforms in the United Kingdom and the European Union, with the UK scoring slightly higher. Facebook, Instagram, Threads, and WhatsApp all rely on the Meta Ad Library as their primary transparency resource and offer broadly similar features across both regions. The main difference is that Meta allows the extraction of political advertising data in the UK, whereas in the EU, political advertising has reportedly been banned to avoid compliance with the TTPA regulation (Votta 2025). As a result, the data export feature is not available through the GUI for any ad type in the EU. Because this criterion is heavily weighted in our methodology, Meta scored comparatively higher in the UK. In the EU, its ad data transparency was classified as Limited, alongside platforms such as YouTube (Google), TikTok, and LinkedIn. Nonetheless, the overall scores for advertising transparency in the EU and the UK remain relatively close.
The artificial distinction between political and non-political ads also reduces access to Meta advertising data in Brazil. There, Meta’s ad transparency is mostly restricted to content labeled as political or related to social issues. By contrast, in the EU and the UK, the Ad Library API provides comprehensive access to data on all ads delivered to users, including both active and inactive ads, with information on advertisers, funding sources, audience demographics (such as age and gender), impressions, temporal data, and moderation actions. In Brazil, however, both the API and the GUI limit structured data extraction to political and social issue ads, while other ads can only be viewed through the interface but not exported. The absence of statutory transparency requirements for all advertisements in Brazil weakens the enforcement of consumer protection, child protection, and electoral regulations, among others.
Google’s approach to advertising transparency illustrates a compliance-driven model that fragments access across jurisdictions. In the EU, the company provides comprehensive programmatic access to data on all active and inactive ads served over the past year, including advertiser information, verification status, moderation flags, and delivery dates, broadly mirroring the Ads Transparency Center interface. As the GUI’s data export is restricted to ads labelled as political—and such ads were banned in the EU following the entry into force of the TTPA—this functionality is no longer available in the region. In the UK, data export features, whether through the GUI or the API, are limited to political advertising, while commercial ads can only be viewed through the interface with more limited information, placing the platform in the Deficient band. In Brazil, Google’s mid-2024 ban on political advertising, motivated by new norms introduced by the country’s Electoral Court, has effectively eliminated structured access to up-to-date ad data, whether via the GUI or the API, while restricting visibility to commercial ads from verified advertisers only, a limitation not present in the EU or UK, placing the platform in the Negligible band.
Across all regions, Google’s structured advertising data lacks ad content, audience demographic breakdowns, and granular spending information. More critically, both programmatic data access and the Ads Transparency Center GUI are constrained by the inability to locate ads through keyword searches, as queries can only be performed based on the identification and selection of advertisers. This design effectively requires researchers to already know who advertised in order to retrieve information, significantly hindering data discovery and raising concerns that the system operates as a form of transparency washing (Zalnieriute 2021).
The TikTok Commercial Content Library is another example of regulatory spillover from which the UK benefits, as, although it was introduced to comply with the DSA’s requirements, it also includes data on ads served to users in the country. While its GUI is publicly accessible to users worldwide, it only includes ads served to users in Europe. API access is restricted to approved researchers, providing structured data on advertiser identity, funding sources, targeting criteria, and delivery dates. However, rate limits are undocumented, spending and engagement data are unavailable, and the interface does not support structured data export. Overall, this places TikTok’s advertising data transparency in the Limited band in the UK and the EU, while remaining unavailable in other regions.
LinkedIn stands out as the only platform to offer global API access to its ad library across all ad types, although significant regional limitations remain. While the API performs more robustly for ads served in the EU—where LinkedIn was classified as Limited—it provides incomplete data for ads delivered outside the region. Notably, the API does not return ad creative content in any jurisdiction, offering only the ad URL alongside basic advertiser and metadata. In addition, both the ad library and the API fail to disclose key information such as ad delivery dates and targeting criteria, although these data points are available for EU-served ads.
Several advertising transparency features available in the UK point to the presence of regulatory spillover effects. Functionalities of the Meta Ad Library originally restricted to ads served in the EU have been extended to the UK, and the TikTok Commercial Content Library is likewise accessible in the country. None of these tools is required under any UK regulatory framework, but the UK is the only market outside the EU to be significantly affected by such DSA-related decisions and measures. However, even this spillover is asymmetrical and only extends to a few platforms, underscoring the need of equivalent domestic guarantees to secure and broaden transparency.
The design of ad repository search functionalities has implications that shape the possible forms of accountability. When platforms require users to search by advertiser name—rather than by ad content, topic, or targeting criteria—the architecture itself becomes a mechanism of opacity. Identifying fraudulent, misleading, or harmful ads becomes effectively impossible without prior knowledge of the advertiser’s identity. The result is an architecture that favors browsing and exploratory use over systematic, data-driven investigation. This limitation is particularly consequential in the context of influence operations, paid disinformation campaigns, and microtargeted political advertising, where the harm lies in the granular dynamics of audience selection and message delivery that are not captured by aggregated data.
The opacity of advertising data has direct consequences for the oversight of electoral campaigns and democratic governance. Without the ability to independently verify the volume, content, targeting, and spending of political ads on social media platforms, electoral authorities, researchers, and civil society cannot meaningfully assess whether campaigns are operating within legal boundaries. This undermines the capacity to ensure fairness and equal opportunity among candidates. In jurisdictions such as Brazil, platforms have responded to regulatory pressure by banning political advertising without extending transparency obligations to advertising more broadly—unlike in the EU, where such bans coexist with comprehensive, regulated transparency requirements. This asymmetry compounds the problem: political ads continue to circulate without oversight, effectively rendering electoral advertising regulation unenforceable.
The lack of standardised advertising data across platforms is a fundamental barrier to effective sector governance. Variations in format, granularity, coverage, and definitions make consistent cross-platform comparison nearly impossible, limiting regulators’ ability to assess the ecosystem holistically and constraining the production of reliable comparative research. Without interoperable standards, platforms operate in isolation, under distinct disclosure practices, undermining efforts to ensure information integrity across the digital environment.
Advertising opacity also undermines both advertisers’ commercial interests and platform credibility. Digital advertising is central to the business models of major social media companies and is projected to reach a global market value of USD 650 billion in 2025 (Precedence Research 2025). Yet transparency and independent oversight remain limited. In the absence of independent verification of ad delivery and performance, investment decisions rely on unauditable, platform-generated metrics—“magical numbers” that are widely cited but fundamentally unverifiable. This erodes budget accountability, distorts competition, and weakens confidence in the online advertising marketplace. When performance claims cannot be independently verified, advertisers are more likely to direct budgets toward ineffective or inflated advertising products.