Twitter Says Its API Can't Be Used For Surveillance, But What Does It Think The FBI's Going To Do With It?

Source: Techdirt.com

Dataminr, the company whose Twitter firehose access has become somewhat of cause celebre on both sides of the privacy fence, is back in the news. After being told it couldn’t sell this access to government agencies for surveillance purposes, Dataminr had to disconnect the CIA from its 500 million tweets-per-day faucet.

Twitter was pretty specific about what this buffed-up API could and could not be used for. The CIA’s surveillance efforts were on the “Don’t” list. This rejection of the CIA’s access was linked to existing Twitter policies — policies often enforced inconsistently or belatedly. What the CIA had access to was public tweets from public accounts — something accessible to anyone on the web, albeit with a better front-end for managing the flow and an API roughly 100x more robust than those made available to the general public.

The question now is how Twitter defines surveillance.

The FBI will soon be able to search a vast repository of public tweets in real time for hints about potential terrorist attacks and other public-safety crises.

The bureau awarded a sole-source contract to Dataminr, a company that allows customers to churn through Twitter’s “firehose,” which includes more than 500 million 140-character messages posted daily. Twitter’s public API only gives users access to about 1 percent of tweets, according to a FedBizOpps posting.

Now, the question is not whether or not the FBI should have access to publicly-available Tweets. It always will have that access, with or without Dataminr’s assistance. The question is whether Twitter believes the FBI is not engaged in the sort of surveillance it disagrees with.

In the context of its Dataminr access, I’m sure the FBI would have preferred to be thought of as a law enforcement agency. Divorced from the API-access context, it has done much in recent years to place itself on the same level as the CIA. It honestly feels it should be given more foreign intelligence gathering powers — more so than the CIA, which has traditionally handled only foreign-facing operations.

Likewise with the NSA. The NSA’s bulk collection orders under Section 215 were obtained in the FBI’s name, with the data going directly to the NSA and the intelligence agency “tipping” an unspecified amount of the haul back to the FBI for further examination.

What the FBI is going to engage in with this access will be a form of surveillance, albeit one with very few privacy implications. Twitter has yet to speak up about the recently-awarded contract. It may never do so. It may believe the FBI is primarily engaged in law enforcement, even though the agency rebranded in the midst of the Snowden leaks, emerging as the “national security” agency it apparently felt it always should have been.

The statement issued by Twitter suggests it’s only the “surveillance” that bothers them, not so much what each government agency seeking access feels its core mission is. The policy says “government or intelligence agenc[ies]” will be forbidden from purchasing access for surveillance purposes and the FBI certainly can’t deny it’s a government agency.

It also shouldn’t matter which hat the FBI wears when attaching the hose. Twitter yanked Geofeedia’s API access after discovering it was selling access to law enforcement agencies all over the US for the purposes of tracking First Amendment-protected activity. Its policies also list “track” and “investigate” as problematic uses of its API — two things the FBI does often.

Given the agency’s long history of engaging in surveillance of protected political activity, it’s not much of a stretch to believe the FBI will use Dataminr’s tools for the same ends. Then again, Dataminr or no Dataminr, the tweets it’s seeking to analyze are already out there where anyone can see them. All the agency is really buying is a hose and a funnel.

Permalink | Comments | Email This Story
Source: Techdirt.com

WordPress Video Lightbox
Cerca