Platforms for user generated content (UGC) play a major role for children and young people nowadays. In such environments service providers usually do not offer any inappropriate content themselves. Instead, they only provide the framework or platform for relevant content that is uploaded by the individual users. And yet many UGC services have implemented tools for the protection of younger users on their platforms, such as user-generated content classification and reporting tools for inappropriate content.
In the context of the MIRACLE project the working group “Child Protection in Social Media” (CHIPSOM) has recently been established. The working group’s objective is to collect the UGC providers’ current approaches in the area of age classifications and labelling, to analyse experiences and outcomes of these approaches and to identify best practice solutions.

The working group acknowledges that there is no technical one-size-fits-all solution due to different innovations, international differences in operating procedures and (often internal) standards, different evolutionary levels of implementation as well as their contexts and options. We started from the perspective that the existing approaches rather need to be strengthened by furthering their potentials through technical interoperability. By standardizing age classification-related meta data of user generated content, the working group aims at improving consumer information and child protection tools on web 2.0 platforms.

Against this background the participants of the kick-off workshop convened on 9th September 2015 to discuss current classification and labelling practices on UGC platforms as well as the potentials of using interoperable classifications.

Key Takeaways of the workshop

  • There’s a lot of classification knowledge stored by user-generated content platforms, but it’s usually locked away.
  • There are relationships between the steps of content classification and content labelling, but usually any form of content classification can use any form of electronic labelling.
  • Current user-generated classification approaches rely on ex-ante uploader self-classification or ex post community classification.
  • Current practices of using the user-generated classification knowledge is limited to internal measures. Public labelling is non-existent.

Find the full workshop minutes here (pdf).