Accountability projects that track deleted tweets from politicians and public officials suffered a critical setback this week when Twitter killed their ability to collect that information. This move comes a few months after the service shut down the U.S. version of Politwoops, the best known of these projects, and extends the ban to some 30 other jurisdictions.

It would be a disappointing move by any platform so popular among politicians, but especially so for Twitter—which as a company has fought vociferously for its users' rights, and made its stance as the "free speech wing of the free speech party" a component of its corporate culture, and a selling point for new sign-ups around the world.

Journalists have drawn parallels between Twitter's decision and Europe's "right to be forgotten," which allows individuals to delist "irrelevant" information about themselves from search engines. Of course, there are some major differences between the two. Most notably, Twitter's policy is issued by a private company, not a court. And while free speech advocates (including EFF) have taken issue with the sweeping implementation of Europe's rules, it differs from Twitter's policy by having notable exceptions for public figures.

More fundamentally, the "right to be forgotten" is at least nominally about privacy. In the case of politicians making public statements, the privacy calculation is totally different, and the need for transparency and public accountability is a major consideration. We acknowledge this calculus in other places—politicians must disclose certain communications under transparency laws and the Freedom of Information Act, for example—but Twitter doesn't appear to have taken it into account.

As Twitter has become increasingly popular among both journalists and politicians, it's been vertically integrated in the realm of public statements. That is, politicians will frequently use the platform to take a stance or react to an issue, and in many cases news reporters—who might have previously called for comment, or quoted from a press statement—will embed the tweet directly.

Embedding a third-party-hosted tweet in a news article has consequences for privacy—it gives Twitter the possibility of tracking the reading habits of users around the web—but also raises concerns about the longevity and historical value of that reporting. Thanks to some forward-thinking engineering, Twitter embeds are currently written to fall back to displaying the text of a tweet, even if it is later deleted or if Twitter is somehow inaccessible, but that fallback has two major limitations.

For one, it will preserve the text of a tweet, but it won't continue to display embedded media—such as the images that the Hillary Clinton and Jeb Bush campaigns have been exchanging. For another, it isn't guaranteed to be the case for newly generated embed codes. Existing embeds will continue to display the text of embedded tweets in perpetuity, but nothing prevents Twitter from re-writing its embed codes to prevent that for the future—which, given this week's development, must be considered as a real possibility.

When politicians make a statement to the press, they necessarily give up control of its use—even if they later regret their words. But when the press serves those statements from a third party under that politician's control, it creates an unprecedented revocation ability.

And the effect on public records laws, too, is unclear. Public communications from state and local elected officials are generally subject to these laws, and have retention requirements. How do deleted tweets fit into this picture? That isn't necessarily Twitter's problem, but it is now blocking transparency organizations from providing a solution.

From a technical perspective, this isn't a new policy. Twitter is enforcing an existing rule in its terms of service, by revoking or threatening to revoke access to the API for organizations that are publicizing deleted tweets. That doesn't make it impossible to track deleted tweets—in fact, the API continues to make notifications of tweet deletions available, so third-party clients can remove those tweets from users' timelines.

Rather, this restriction is disproportionately felt by the transparency groups that are publicizing their work and have a good reason to remain on good terms with Twitter. In other words, advertisers and individuals that are monitoring deleted tweets for privacy-invasive reasons aren't technically restricted from doing so, but organizations working in the public interest are blocked by policy from “surfacing” them.

Again, it's worth noting that Twitter has gone to great lengths to defend free speech in the past. And to be completely clear, nobody is arguing that the company is under some kind of legal obligation to make deleted tweets available. But especially coming on the heels of Twitter's compliance with a controversial and likely bogus takedown related to the Ashley Madison data breach, it's disappointing. Twitter may be sparing politicians some embarrassment, but at a cost to the public interest.

Related Issues