The recent debacle over the TwitPic terms of service has illustrated just how vulnerable the Web is to TOS abuse, whether intentional or accidental.
One of the points that is often overlooked in the saga is that TwitPic first changed its TOS on May 4th, not to have anyone notice the change until May 10th, almost a week. Considering how egregious the change was and how popular TwitPic is as a service, it’s worrisome that it wasn’t caught earlier.
However, the truth is that most people don’t read the terms they agree to and, when they do, they likely don’t understand what they are agreeing to due to the language used and the sheer size of many of the TOS pages.
TwitPic may have illustrated the danger, but the problem is larger than one site or service. Instead, the problem is one faced by almost the whole of the Web and, as such, needs to be addressed accordingly. However, fixing it won’t be an easy task and involves trying to address multiple issues at once, some of which don’t seem to have any simple answers.
Still, with an issue this important, and one that goes well beyond copyright law, it’s important to at least make an effort and start generating ideas to solve these issues.
The 3 Problems With Nearly Every TOS
No matter how little a company attempts to grab rights or be sneaky with their terms of service, the nature of having a long, legalese-filled contract between the user and a service they want to use creates a slew of problems that are difficult to overcome.
To make things easier, we’re going to break up the problems into three key issues that would need to be addressed in order to make the TOS signing process both fair and orderly.
- Most people don’t read the terms of service because they are too long and mostly the same.
- Those who do often don’t understand what they are agreeing to due to the language.
- Even if one does read and understand the TOS, the terms often change, usually without notice.
These issues create an environment where nearly everyone, lawyers and laypeople alike (myself included) rarely give the proper amount of attention or understanding to the terms they agree to.
So how do we address these issues? Lets take a look at them one at a time and try to understand what, if anything, can be done.
Problem 1: TOSes are Too Long
Apple’s iTunes is famous for it’s unbearably long terms of service, clocking in at over 15,000 words (Note: An average article on this site is between 800 and 1,200 words), but it is far from the only service with long a long TOS.
Facebook, for example has over 4,000 words in its main TOS, not counting its various other TOS pages. However, Facebook merely ties Google, which has almost the same number of words in its main TOS (once again, not counting the TOSes for many of its various products).
To make matters worse, much of the content in these terms are stock. If you put either TOS into a standard plagiarism checker, you’ll find matching text in dozens, if not hundreds, of other sites. But even if the exact text isn’t the same, the meaning often is.
There needs to be a way to cut through the stock elements of a TOS and parse it down to the important “need to know” information.
On that front, Java Cool Software created a produced called EULAlyzer, which is designed to do something similar but for EULAs that come with software, not Web service terms of service. Though my testing found it to be a poor fit for this type of issue last year, it certainly is a proof of concept that something like this could be made and may help raise red flags at appropriate times.
While not a complete solution, such a tool, especially if it were Web-based, could help cut through the cruft and identify key issues that the user needs to be aware of before clicking “agree”.
Problem 2: TOSes are Too Difficult to Understand
Even if your average TOS were reduced in length enough to make the average user sit through it, understanding what it meant would still be a challenge for most.
The reason is that most terms of service are written by attorneys for attorneys. They are written in legalese in the event of a legal dispute. Unfortunately, this means that the confusing language and terminology has to be in there for the TOS to do its job, however, that doesn’t help users make heads or tails of what they are agreeing to.
A more robust solution to this problem would call upon service providers to actually make their terms more accessible. For example, following a Creative Commons model of making the licenses machine, human and lawyer-readable at once. While that would be difficult, especially considering all the areas that a TOS has to cover, but creating this system would enable machines to parse terms easily, humans to read them and lawyers to feel confident in the protection they offer.
Another solution, and another way the Creative Common model may be able to help out, is to create a series of “stock” licenses that are explained on a third-party site. Then hosts and service providers would be able use those licenses and add their own terms as needed.
For example a host could say that: “Usage of our site is provided under [License Name] and is subject to those terms as well as the ones below.”
If the main terms were reasonably well explained, the rest, most likely would be shorter and easier to understand as well, making the TOS both much shorter and easier to read.
All in all, companies will be reluctant to go down either of these paths. The first requires additional effort and work, all for something few feel is important. The second would save costs in drafting a TOS, but requires turning over a key legal document to a third-party. In short, neither solution is likely to be adopted, though they may still be ideas worth exploring.
Problem 3: TOSes Change Routinely
Of all the problems on the list, this is the most “solved” one. Simply put, a terms of service is not likely to stay the same very long as they are updated very routinely. However, few users have the time, energy or even the awareness to follow along with the changes and make sure they still agree.
On that front, the EFF launched TOSBack, a terms of service tracker that monitors TOSes for changes and reports on what’s different. Unfortunately, the system is limited to just 56 sites and isn’t always perfect. For example, the PayPal link doesn’t appear to be working.
What may be more practical is existing change detection systems such as ChangeDetection.com, which monitor webpages for alterations and report back on any differences it spots.
However, both of these systems are prone to false positives. If a company as much as fixes a typo in their TOS, an alert is triggered. This was a problem I had noticed back in 2009, when the service was launched.
While these settings can be tweaked to reduce false positives, obviously the better system is for companies to notify users of any important TOS changes. However, since most terms of service don’t require companies to do that, few do and selling the idea is going to be an uphill battle. Even good companies don’t want to risk turning customers away with a greater (mis)understanding of their TOS.
Obviously, this isn’t going to be a problem that we can fix overnight. The Web has been built on the current TOS system but, through a combination of new laws and a tidal wave of new services, the TOSes have simply grown too long, too confusing and too numerous to be of much practical use to users.
Fixing this won’t be easy and will require the cooperation of many different parties, but if we can at least improve the situation, we’ll have a much more educated and much more sane Web, one where incidents like the TwitPic case are incredibly rare
It may be a dream, but it is a nice one to have.