As the European Parliament prepares to vote on the controversial “Chat Control” regulation, new research from Incogni’s 2025 Social Media Privacy Ranking highlights how most major social platforms already operate on extensive user data collection – raising questions about what privacy remains to protect.
The study analyzed 15 of the world’s most popular social media platforms across 14 privacy criteria, including data sharing, AI use, and user control. It found that:
- 73% of platforms admit to using personal data for AI model training or algorithmic personalization.
- Platforms such as Facebook, Instagram, and TikTok collect up to 18 distinct data types per user, including identifiers, behavioral metrics, and location data.
- Meta’s platforms (Facebook, Instagram, WhatsApp, Messenger) and TikTok rank among the most privacy-invasive, with penalties across almost every evaluation category.
- Even relatively privacy-friendly platforms like Discord rely on third-party service providers – a recent incident involving one such provider demonstrated how indirect exposure can compromise user data despite strong in-platform policies.
- Several major platforms, including Meta’s and LinkedIn, indicate that they may collect sensitive data such as sexual orientation, race or ethnicity, and health information.
“The Chat Control proposal raises serious questions about privacy and encryption,” said Darius Belejevas, Head of Incogni. “Our research shows that most major platforms already process personal data for algorithmic and AI purposes – meaning the digital environment is already deeply exposed. Adding new layers of scanning on top of that only amplifies the risks.”
Privacy advocates warn that the EU’s proposal, which would require client-side scanning of private messages to detect illegal content, could erode encryption and expand monitoring powers. Incogni’s findings show that large-scale data collection is already the norm across social platforms… long before any new rules take effect.
Transparency and Government Access
Beyond data collection and AI use, Incogni’s research also examined how social platforms handle government requests for user data and how transparent they are with both regulators and users.
In the Transparency category of the 2025 Social Media Privacy Ranking, platforms were assessed based on the frequency of government disclosures worldwide and the accessibility of privacy information available to the public.
- Snapchat disclosed user data in response to 82% of government requests – the highest among analyzed platforms.
- Meta’s products followed closely, disclosing data in 78% of cases, while Discord disclosed in 77.4%.
- X (formerly Twitter) shared data in 51% of requests, followed by LinkedIn (53%) and Pinterest (54%).
- Telegram, Quora, and Twitch provided no public data on this measure and were penalized for lack of transparency.
Platforms with detailed privacy centers performed best, but the high disclosure rates across major networks demonstrate that user information is already frequently shared with authorities under existing frameworks.
About the study
The 2025 Social Media Privacy Ranking by Incogni’s Research Lab evaluates leading platforms on the transparency and scope of their data practices. The analysis covers data collected, shared, and processed for algorithmic and AI purposes, based on publicly available information accessible to ordinary users.
About Incogni
Incogni helps (in the company’s words) “individuals protect their privacy by automating personal data removal requests from data brokers and other third-party databases.” Its research lab investigates global data collection trends and emerging privacy risks.
