4 Comments

Summary:

Do you know what kind of information your Twitter apps are collecting? According to Mike Champion, VP of engineering at app directory Oneforty, the way that Twitter handles user permissions means that you could be giving those apps access to plenty — including your private messages.

3578914701_5f51c49174_z

Do you know what kind of information your Twitter apps are collecting about you? If you’re like me, you probably just click on the various permission screens to allow your iPhone app or your desktop client to collect and publish messages for you, and don’t think about it after that. However, the way that Twitter handles user permissions means you could be giving those apps access to plenty of information — including all of your DMs, or private messages — as Mike Champion, the VP of engineering at Twitter app directory Oneforty, notes in a recent blog post.

Champion points out that, because of the way Twitter has implemented OAuth, the open user-identification standard it recently adopted, applications that request your permission to connect to the network effectively only have two options: one allows them “read-only” access through your account — that is, the ability to pull Twitter messages in and display them — and the other gives them “read-write” access, which means they can not only read messages, but can also publish them to the network on your behalf. Since most app developers want to do more than just read messages, the majority of apps request full read-write access. The risks in doing so, Champion says, include:

  • DM privacy. As Champion notes, many people use direct messages as a private communication channel, but any app that has read-write access can read all of your DMs, something that might be a concern for certain users.
  • Spam and hackers. If you authorize full read-write access to a malicious app, or someone gets hold of your “token” that gives an app permission, they could send malicious links through your account, or even delete all your messages.
  • Pushy apps. This would include services that auto-publish to Twitter without asking you, or without making it obvious that they are going to do so, such as Twifficiency, which recently caused some backlash by auto-tweeting on a user’s account when they signed up for the service.

As Twitter continues to become a real-time messaging layer and news-delivery platform for the web (with all that implies), users are likely to start paying even closer attention to how their data is distributed by the network. Many people have gotten used to treating Twitter as another communication service, just like email and instant messaging, but Champion’s post is a welcome reminder that we should be aware of what we are allowing apps to do on our behalf.

Twitter may also want to think about how to allow a little more customization in its permissions, rather than just giving developers a binary choice between read-only and full write access. Earlier this year, Facebook made changes to the way that applications handle permissions from users for various activities, which requires them to ask users each time they want to use a piece of personal information.

In response to a request for comment, a Twitter spokesman said that the move to OAuth “is a big step forward for user security” because a user’s name and password are not stored by the app, but that the service is “always looking to improve user security and privacy” and is “actively working toward developing better solutions.”

Related content from GigaOM Pro (sub req’d): Why New Net Companies Must Shoulder More Responsibility

Post and thumbnail photos courtesy of Flickr user Marina Montoya

You’re subscribed! If you like, you can update your settings

  1. Ars Technica recently posted a great article on how screwed-up Twitter’s implementation of OAuth is:
    http://arstechnica.com/security/guides/2010/09/twitter-a-case-study-on-how-to-do-oauth-wrong.ars

    But as for DMs, I don’t see that as an unexpected security risk. I mean, don’t most Twitter clients have the ability to read/send DMs amongst their features? And the security issues with read/write access (apart from the OAuth issues) would apply to any client that needs read/write access, including Facebook (and its private message feature). If the app can let you post status updates and read DMs, it will have the ability to post updates on its own or read DMs as well.

  2. Like most people, I don’t even think about what information Twitter is collecting on me when I go through my daily routine of Tweeting and maintaining the account. @jgwentworth

  3. Coworking in Florida Monday, September 20, 2010

    Twitter DMs have been a nuisance for me. I wonder if Twitter could create an on/off switch for this module, or certain filters.

    Most DMs we receive are either an auto-DM, a useless-solicitation DM, or a DM as a result of a hacked account.

    Here is a blog post I wrote on this previously:

    http://blog.officedivvy.com/2010/07/this-is-why-twitter-direct-messages-dm.html

    Ky

  4. How Twitter DMs can actually be seen publicly by developers and 3rd parties Monday, October 11, 2010

    [...] are lots of technical explanations on who can see your Twitter DMs which you can read here and here but essentially to sum it up anybody who is devious enough could see your DMs. Basically [...]

Comments have been disabled for this post