TGEU and IJSC publish report on trans people’s experiences of online hate speech

TGEU — Trans Europe and Central Asia and the Institute for Journalism and Social Change (IJSC) carried out a study to investigate trans people’s experiences of online hate speech to understand the extent of hate speech against trans people and how this type of content may get flagged. The study was based on a survey of trans content creators and a social media content analysis in English on YouTube.

The study found that trans social media users and content creators across Europe experience hate speech, coordinated abuse, misgendering, and platform bias at alarmingly high rates. To make online spaces safer for trans people, it is essential to understand:

• the frequency and scale of hate speech targeting trans social media users 
• the types of hate speech that occur
• where it comes from and how it spreads online 
• and how, or whether, social media platforms respond

30 trans people in 16 countries in the Council of Europe region took part in our survey. Covering countries such as Azerbaijan, Czech Republic, Finland, France, Georgia, Germany, Ireland, Italy, Latvia, Netherlands, Northern Cyprus, Norway, Poland, Romania, Switzerland, Ukraine. 

Key findings

• All trans content creators who responded to the survey reported experiencing hate speech. 35% of respondents reported experiencing hate speech daily, while 53% said they received it weekly.
• Trans content creators surveyed reported hate speech across several platforms, including Facebook, Instagram, TikTok, Youtube, X, Reddit, BlueSky, Mastodon, and language-specific platforms such as Odnoklassniki.
• Abuse was also reported in response to personal content, including videos of walking in public, selfies, and daily updates. This illustrates how even ordinary forms of online self-expression and visibility as a trans person can attract targeted harassment. The responses also reflect how forms of hate speech vary, are often repeated, and are explicitly aimed at delegitimising trans identities, such as consistently referring to the person with their deadname or claiming they had no right to life. 
• An overwhelming majority of survey respondents selected multiple categories of harm. The most common forms included:
  • Denying their existence as trans people, reported by 67% of respondents
  • Homophobic or transphobic slurs (63%)
  • Misgendering or deadnaming (57%)
  • Threats of physical harm (57%)
  • Death threats (50%)
  • Rape or sexual abuse threats (37%)
  • Threats to family or friends (10%)

YouTube content analysis results

The content analysis used two approaches. First, we conducted a programmatic test on a sample of 1,628 English-language comments taken from 11 YouTube videos published between July 2023 and 2025. Using an extensive list of 168 keywords and phrases, informed by the 2025 survey of trans content creators across Europe, we tested the effectiveness of programmatic detection of hateful and discriminatory language.

Secondly, we adopted a mixed-methods approach, which included manual coding for comparison. This method identified approximately twice as many harmful comments as the programmatic test. Overall, the flagged comments accounted for around 10% of all comments analysed. Most of the hateful and discriminatory comments identified by both methods appeared beneath videos published by mainstream media outlets.

These findings point to multiple layers of accountability. Both the platform itself (YouTube) and the publishers responsible for the videos, in this case the media outlets, have a duty to protect trans people in digital spaces. Additionally, questions must also be raised about why such content has been allowed to remain online, particularly in light of the Digital Services Act (DSA), which came into force across the European Union in early 2024.

The research also highlights the limitations of automated moderation systems. The programmatic test nearly always flagged fewer instances of hate than the manual efforts, whether related to hateful or discriminatory content, dog whistles, or threats. This reinforces long-standing concerns raised by civil society organisations:  automated systems for detecting online hate speech remain inadequate and often fail to capture the specific forms of abuse experienced by trans communities online.

More info

• Submission to the MSI-eSEC Public Consultation on the Draft CM Recommendation on Online Safety and Empowerment of Users and Content Creators
• Code used for extracting comments from all of the videos
• Code used for extracting video publication dates and video count
• Code used for identifying keywords and phrases in comments

Your support makes change possible

We work across Europe and Central Asia to advance trans rights, build strong communities, and drive change through research, advocacy, and community-building.

Your donation helps us continue this vital work — defending trans lives, amplifying trans voices, and advocating for justice every day.