Blog, Uncategorised

The ISP Porn Filter or why you cannot have a Unicorn

After the government have gone to so much effort to pester ISP’s into creating a default filter in order to “protect children” from the evils of the internet (mostly porn), three ISP’s have started implementing it. So what is the result? Well, unsurprisingly really, it turns out the geeks were right and the filter is a deeply flawed idea which doesn’t fix the problem it is meant to, and in fact creates new problems.

It turns out that the ISP filter does not block all porn/harmful content, but it does block innocent content. Which is helpful, and utterly predictable.

The short of all of this is that if you have children and want them to be kept safe online then even if you choose to use this filter, you will still need to install another filter and supervise your children. This filter will not actually make your life any easier, and if you choose to rely on it alone without bothering to install an additional device based filter then it may leave your children far more exposed.

This isn’t because it’s a badly designed filter, I have no doubt they are using the most up to date technology and it’s probably very very good. But the technology is limited by what is technically possible at this moment in time, and no amount of effort on the part of the geeks is going to change that. They can only work with what they’ve got, and the internet is a strange beast that was never intended to do half the stuff it does do, so the tech that allows it to do so much is mostly a bit of a bodge job – a very impressive bodge job, but still, the fact is we are doing things with the technology that the technology was never designed for.

Rather than take over this whole blog with multiple posts about this I have previously created a separate blog dedicated to the filter and it’s flaws, so apologies if this seems a bit cheeky linking to another of my own blogs. First things first, this is a long post on the technical limitations that dictate what computers can actually do when it comes to filtering and the pros and cons of the different approaches. I’ve tried to simplify it as much as possible, but happy to try and explain further if anyone wants to comment and ask.

The basics are that while you or I could look at a site and say, “yep, that’s porn” or “no, that’s sex education”, a computer cannot necessarily distinguish between the two. Sounds really silly, I know, I mean how on earth could you confuse sex ed for porn?? Well if you look at what data the computer can actually work with it’s actually quite easy. A computer can only use the information it is given, and it can only process certain types of data. So while we can ask, “is this designed to arouse?” and answer it, a computer has no idea at all how to quantify that, let alone answer it. If you want to get a computer to answer a question you need to give it information like keywords, and sex ed sites and porn sites share an awful lot of keywords. Which leaves you with the dilemma of, do you block all sites with those keywords as the porn sites use them? Or do you allow them as the sex ed sites use them? Either way you can see there are flaws, we end up with a filter that is too lax or too strict. This is, of course, ignoring that if someone wanted a porn site to get around a filter they could just change the keywords they used.

Of course you could bypass computers and get humans to decide which sites are ok and which aren’t, this is a much more efficient method in terms of blocking porn sites. However there are so many new websites added every single day that it would impossible to check them all which means you’d need to manage it one of two ways. A black list or a white list.

A black list would mean allowing all websites until they are manually added to a blocked list. As you can imagine that will let a lot of porn through, a hell of a lot of porn.

A white list would mean blocking all websites until they are manually added to an allowed list. A mammoth task which means a lot of innocent sites are blocked until someone can get around to checking them, I’ve heard that it’s possible that one of the filters currently implemented is operating a white list, and now has a huge backlog of people wanting to know why their site is blocked and to get it unblocked. In the meantime people are unable to access things like support websites, informational websites, and some businesses are seeing that their customers cannot access them (not great in our current economic climate). And who knows how long it will take to clear that backlog.

Then there’s the inflexibility of an ISP filter, as soon as you have more than one person in a household with different internet safety needs it all falls apart.

I know it’s frustrating being told that there is no easy solution to make the internet safer for children, it’d be great if there were, but we need to be realistic. Just saying there “should be” doesn’t make it so, we’ve got to work with what we’ve got. And right now what we have are a lot of different programs that can be installed on devices and customised to suit individual families needs, they aren’t ideal, and still require a level of supervision. But they are less flawed than the ISP filter, and still the safer bet if you really want children kept safe. Their biggest flaw is the number of people using them, so our efforts should be focused on getting more people using them, rather than putting people off by making them think the ISP filter will do the same job.

Oh and the unicorn reference? Here you go

Leave a Reply

Your email address will not be published. Required fields are marked *