Computers, Privacy & the Constitution
DRAFT - Ready for comment.

How Your Tweets Are Forever Searchable

-- By JulianBaez - 16 Apr 2010

What You Thought

Your Twitter account is public by default. Moments after a high profile vote on healthcare reform you punch out a reaction in a minute. “Democrats are doubling the deficit and enacting a government takeover of #healthcare,” or “Republican #HCR filibuster = 30 million sentenced to death or bankruptcy.” Your followers, all personal friends, see the tweet. In 12 hours, it’s no longer visible on their front pages, other tweets have pushed yours off their front pages.

A stranger searches the hashtag #HCR or #healthcare. If they search within two weeks, your tweet appears. If they search after two weeks, your tweet is no longer indexed and searchable. Your momentary reaction is only available to those who look specifically at your twitter.

Or so we thought.

What’s True

In the summer of 2009, the Library of Congress stated that it intended to compile all tweets related to Sotomayor’s Supreme Court nomination, via Twitter no less. This was part of the Library’s Web Archiving plan to preserve primary source materials. On April 14th 2010, Twitter announced it would donate its entire public tweet archive to the Library: “Only after a six-month delay can the Tweets will be used [sic] for internal library use, for non-commercial research, public display by the library itself, and preservation.”

But what use is information in 2010 without means to quickly search it? Google Replay allows users to search tweets available on a specific date and time. Eventually, “it will reach back to the very first tweets.” Your instantaneous reaction will become instantaneously searchable ad infinitum.

We still don’t know whether user’s profile information, followers, geo-locational data and deleted tweets will be available. Will users be classified like books?

The Law


Intellectual property rights in blogs and tweets are a newly emerging legal question courts have yet to address. The blogosphere’s prevailing opinion is the majority of tweets are not copyrightable because they are not original works of authorship, do not reach the requisite level of creativity, and are simply too short to register. Unlike similarly succinct haikus, most tweets are simply statements of fact, making them not copyrightable.

Twitter’s terms of use relinquish any intellectual property rights in material posted and claim the copyright remains in the hands of the posting user. Therefore, Twitter would not claim it held the copyright. The blogosphere consensus maintains that about 90% of all tweets are non-copyrightable statements of facts and references to other facts. Although the specific way of expressing facts has copyrightable potential, 140 characters and common knowledge ideas significantly limit the opportunity for original expressions. Most bloggers also believe the chances of individuals pursuing copyright infringement claims of their tweets are slim to none.

You probably do not own your tweets; you probably do not care.


There’s also privacy concerns which can be divided into two parts, tweets before and after Twitter’s announcement that all past and future tweets will be searchable.

There is one big difference between the two; tweeters could argue they had a reasonable expectation of privacy before the announcement. Twitter’s public operating procedure had been that all tweets were no longer searchable after two weeks. While researchers could individually look up your individual twitter account, they could not come across your tweets through a general search after two weeks. A user could reasonably expect their tweets would not be as prominently displayed before the announcement, even if the tweets were still public.

Whether we are discussing tweets made before or after the announcement, current privacy law likely doesn’t cover public tweets. However, there are cases which hold that the fact the privacy one expects in a given setting is not complete or absolute does not render the expectation unreasonable as a matter of law (Sanders v. ABC (Cal. 1999)). Holdings like this leave open the possibility for tweeters to claim tort of intrusion even in a public place.

Although politically unlikely for multiple reasons, a statutory safeguard for tweeters would provide certain protection. There is recent legislative precedent for protecting privacy in public places. However, the Video Voyeurism Prevention Act of 2004 only protects individuals in public places when they reasonably believe their private areas would not be visible to the public. Also tweeters are a far less sympathetic group than victims of voyeurism. Finally, a tweet protection statute would have to be clear that a reasonable person has an expectation of privacy after a specified time following a public tweet.

The Consequences Of A Public Tweet Database

My prediction is that any controversy over this will mirror closely the controversy surrounding Facebook’s introduction of the mini-feed feature. A small group of users will complain about the feature as it is fully introduced and its repercussions are felt. This will result in a public apology and the introduction of new copyright and privacy controls for users to opt in to. In this case, Twitter will likely allow users to be entirely private, public, or non-archived public. This will somewhat alleviate concerns but less users will opt in to the additional copyright and privacy controls over time. Finally, these controls will be slowly relaxed until they are unrecognizable or entirely removed. The frog will boil slowly.

The final result will more or less mirror the initial concerns. However, users who did not initially give up the service or adopted the service afterwards, will barely be aware of the privacy and copyright issues. Basically, despite whatever ruckus you hear in the coming weeks and months, Google Replay and the Library of Congress now have all your tweets.

Your momentary reaction will be forever searchable.

This is an interesting essay, valuable for what it shows yet again about the dangers of unfree services. There wasn't any reason to centralize microblogging anymore than there would have been for centralizing blogging, and trying to get everybody in the world to put their blogs in the same place, so that server operator would have all the logs to analyze and sell. Twitter like Facebook is an architectural pathology, and what you're describing is a part of the pathology. Of course, the real point is that by making their data public they are acknowledging that Google already has the opportunity to do so, which means they can't sell it. And Google has the advantage that although they allow everyone else to search the stream, they alone can combine that whole stream with everything else they also have, to create a richer combination than anyone else could make.

The solution isn't to be found in law, but in technology. We have to give people a way to do better than Twitter for themselves, while also maintaining their privacy, and without making them go cold turkey from the services they already have. I've explained recently why this is what we need and how it could work, and we are now seeing some really important technical developments, which are getting some media attention. I don't think the problem of Twitter is going to last much longer.

Hi Julian, interesting work.

Just a comment that supplements your concerns about the privacy part of a public tweet database: privacy isn't binary (it's not all or nothing) and it encompasses a right of self-informational determination. For example, the fact that a person shares private information on her Facebook page, which is accessible to 100 friends, doesn’t mean that she wants to share it with the world. I don't see how a public searchable tweet database will respect this right, unless the users explicitly and specifically provide their consent for such use

-- NikolaosVolanis - 26 Apr 2010

Hi Julian,

I was very interested in this topic as well when I heard about it. The reasonable expectation to privacy regarding twitter is a rough sell in my opinion. I think the fact that they are knowingly searchable for 2 weeks eviscerates much of the idea that there is a reasonable expectation to privacy. With that being said, I still don't agree with the methods these social networking sites deal with initial privacy settings. The very fact that you have to opt-in to sufficient settings to acquire privacy offends the entire notion of privacy to begin with. Google got a lot of flack for their implementation of Google Buzz, but like you said, Facebook operates in a very similar fashion when releasing "new features," yet most of the privacy concerns are allowed without nearly as much fanfare. I don't get it. Maybe FB is just that ingrained into our society?

In any case, I agree with Professor Moglen that Twitter's death is imminent. I've always thought of Twitter as a rough and dirty way to do what it tries to do. There has to be a better way to do this type of stuff.

-- EdwardBontkowski - 14 May 2010



Webs Webs

r8 - 17 Jan 2012 - 17:48:27 - IanSullivan
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM