Web content filters are inherently broken

A news website I frequently visit is blocked by my cellular network provider’s content filter, and it was also blocked by my previous provider. The website is ostensibly a front-end to another website I view, changing the list of stories in a way that better suits how I want to visit them. That other website isn’t blocked by the content filter.

The page explaining how the content filter operates says that filtered content is “classified as ‘18-rated’ or ‘unclassified’ against a framework developed by the Independent Mobile Classification Body”. The Independent Mobile Classification Body says:

“Content that falls outside of the IMCB’s remit and Classification Framework includes content accessed via the Internet or WAP (where the mobile operator is providing only connectivity). This means that if you were attempting to download a website to your mobile and this content was blocked, this is outside of the IMCB Classification Framework – put very simply, the IMCB Classification Framework applies only to commercial content that is provided specifically for mobiles.'

If connectivity to a website is blocked on a mobile phone, consumers should contact their mobile operator.”

So the IMCB didn’t block the site directly, but it provided the framework for EE to evaluate the website that was blocked, despite the IMCB saying explicitly that the framework’s remit does not cover the site. The framework text was written in 2005.

On top of that, EE describe the content filter has three modes — strict, moderate and off — and two classification criteria — “18-rated” and unclassified.

Only the strict mode is described and only the “18-rated” criteria applies to it. I don’t know what moderate does or what unclassified content is. It clearly doesn’t mean truly unclassified, but something else. The whole thing is incredibly opaque when it should be incredibly transparent, so I can make a truly informed decision about switching it off or changing it to moderate.

There’s also no direct way for me to query what’s on the filter list, request the website be removed from the filter list, or see how the ‘moderate’ mode works.

When I try and turn off the filter, I’m only presented with on and off as choices, and it says it only affects “social networking and 18-rated material”.

So:

  • EE and GiffGaff block one site but allow at least one other showing the same data, just in a different way
  • EE say the 7-8 year old IMCB framework helped them classify the site
  • The IMCB say the framework shouldn’t have applied to the site at all
  • The EE filter description is opaque, incomplete and inconsistently presented
  • The EE filter controls don’t map to the description on their Help site
  • You can’t find out what’s on the list unless you visit something filtered by it

To top it all off, as well as being able to access the original website I’m using a different front end to, scores of adult content remains unfiltered when the filter is turned on and it’s trivial to discover.

The inherently broken bit I alude to in the post title is that fact: it’s impossible, technologically speaking, to filter everything that fits a classification as broad a “18-rated”. To even attempt it makes no apparent sense, other than to appeal to misguided consumer bodies that think its important. Access to adult content will always be possible, filter or not.

Education is what’ll stop young people accessing content unsuitable for their age, along with parental supervision to make sure that as it becomes more suitable, it’s accessed safely and not to the young person’s detriment.