← Latest papers
💻 computer science

Disarranged Harmonization of Transparency Reporting by Social Media Platforms Under the Digital Services Act

This paper presents the first systematic evaluation of transparency reporting data quality for the eight largest EU social media platforms under the Digital Services Act, revealing that despite new harmonization regulations, significant issues persist regarding data formatting, timeliness, consistency, completeness, and interoperability across platforms.

Original authors: Amaury Trujillo, Benedetta Tessa, Stefano Cresci

Published 2026-05-19
📖 4 min read☕ Coffee break read

Original authors: Amaury Trujillo, Benedetta Tessa, Stefano Cresci

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the European Union (EU) decided to build a giant, public "glass house" to watch how social media giants (like Facebook, TikTok, and YouTube) police their own content. The goal was to make sure these companies are honest about what they delete, why they delete it, and how they do it. To make this fair, the EU introduced a new rulebook called the Digital Services Act (DSA). This rulebook demanded that all the biggest platforms submit their reports in the exact same format, like students all turning in homework on the same type of lined paper with the same font.

This paper is like a teacher grading the first batch of homework from eight of the world's biggest social media companies after this new rulebook started. The researchers asked: "Did the students actually follow the instructions, and do their answers match up?"

Here is what they found, explained simply:

1. The "Homework" Was Messy

Even though the EU gave everyone the same template, the "students" (the platforms) didn't all follow the rules perfectly.

  • The Formatting: Some companies used the wrong font, some filled out the wrong boxes, and some tried to hand in their homework in a completely different language than requested.
  • The Timing: The rule said, "Tell us immediately when you delete something." Instead, some platforms waited weeks or even months to submit their reports. It's like a student waiting until the end of the semester to turn in a daily diary.
  • The Numbers: Some companies reported numbers that didn't make sense. For example, one platform claimed to have more "active users" in a small country than the entire population of that country plus all the tourists and businesses there. It was like a restaurant claiming to serve 1,000 meals a day in a town that only has 500 people.

2. The "Two-Story" Problem (Inconsistency)

The EU asked for two types of reports:

  1. The Detailed Log (TDB): A list of every single piece of content deleted, like a receipt for every item bought.
  2. The Summary Report (TR): A monthly summary, like a monthly bank statement.

The researchers found that for many companies, the receipts and the bank statements didn't match.

  • The "Ghost" Actions: Some companies said in their summary, "We deleted 1 million things on our own," but when you looked at the detailed logs, those 1 million actions were missing.
  • The "Robot" Confusion: Some companies claimed in their summary that they used robots (AI) to delete most content, but the detailed logs showed that almost all of it was done by humans. It's like a chef telling you they used a machine to chop the vegetables, but the kitchen logs show a human with a knife did it all.

3. The "Translation" Errors

The EU asked companies to break down their data by language (e.g., how many posts were deleted in French vs. German).

  • Some companies ignored this and gave the data by country instead.
  • Others used the wrong codes (writing "Ireland" instead of "Irish").
  • One company (X, formerly Twitter) had a strange issue where they claimed they deleted almost no content on their own, but their logs showed they were deleting thousands of things every day.

4. The "Secret Teams" Theory

The researchers noticed something strange: For some platforms, the "Summary Report" looked very professional and accurate, while the "Detailed Log" was a complete mess.

  • The Metaphor: It's as if the company has two different teams. One team is the "PR Team" that writes the nice, clean summary for the public. The other team is the "Ops Team" that actually does the deleting and keeps a messy, disorganized notebook. These two teams aren't talking to each other, so the numbers don't match.

The Bottom Line

The paper concludes that while the EU tried to build a standardized system to make social media transparent, the system is currently broken in practice.

  • The data is often late, formatted incorrectly, or contradictory.
  • You cannot easily compare one platform to another because they are all playing by slightly different rules, even though they were told to play by the same ones.
  • The "glass house" is supposed to let us see clearly, but right now, the windows are foggy, and some of the glass is cracked.

The authors suggest that to fix this, the EU needs to stop just giving out templates and start automatically checking the homework before it's accepted, and they need to give clearer instructions on what the words actually mean (like defining exactly what counts as "automated" vs. "human"). Until then, we can't fully trust the numbers these companies are giving us.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →