The FBI recently signed a deal that will give it access to monitor Twitter’s “firehose” — every tweet posted publicly each day, roughly 500 million of them.
The move has raised privacy concerns and claims that Twitter is not being consistent in enforcing its ban on data being used for surveillance purposes.
“It’s deeply frightening that the FBI is about to access, and may already have access to, this very powerful tool,” said Kade Crockford, from the American Civil Liberties Union.
The FBI, like anyone else, has been able to track some publicly-posted tweets all along, but Twitter’s public-facing interface provides users access to only a small slice of tweets at a time, while beneath the surface millions of tweets fly back and forth.
Access to the far more comprehensive firehose data stream is restricted by the social media company.
The FBI recently agreed to a deal with the company Dataminr, which already has firehose access.
In a summary of the deal, the FBI said the deal will allow the agency to “directly access the full firehose” and to “search the complete Twitter firehose, in near real-time, using customizable filters.”
The deal will also allow the FBI to change those filters to “reflect changes in investigative priorities,” according to the document, which was posted this month on a federal website that details government contracts with businesses.
“The FBI has a need to obtain information about relevant breaking news and events in real-time,” the document said. “Twitter is a platform where news first breaks on terrorist attacks, military actions, epidemiological events, and natural disasters, among other topics.”
In a separate document in which the FBI requested that companies bid on providing it with a tool to access the Twitter firehose, the agency said it wanted access to further “its law enforcement and intelligence missions.”
Critics said the FBI’s plan appeared at odds with Twitter’s ban on its services being used for surveillance, and how the company has handled such cases in the past. But Twitter said that’s not the case.
“A narrowly tailored news alert product is available to some first responders, like the FBI,” company spokesman Nu Wexler said in an e-mail to the Globe Wednesday. “Our position on surveillance use cases is clear and those use cases are strictly prohibited.”
Dataminr also defended the deal in a statement, saying, “Dataminr is not a product that enables surveillance.”
“Dataminr offers breaking news alerting products that help over 300 news organizations and hundreds of investment firms & corporations learn about breaking news and breaking events before mainstream coverage,” the statement said.
“A limited version of our breaking news alerting product is also available to the FBI, ensuring that they are also alerted to the same emerging stories and events as early as possible,” the statement added. “Knowing about breaking news quickly is key for ensuring public safety and the fastest emergency response.”
An FBI spokesman declined to comment.
Twitter recently has blocked law enforcement agencies from using its data for surveillance and publicly emphasized its ban on the practice, including in a blog post on Tuesday.
“We prohibit developers ... from allowing law enforcement — or any other entity — to use Twitter data for surveillance purposes. Period,” the post said. “If developers violate our policies, we will take appropriate action, which can include suspension and termination of access.”
In October, the ACLU revealed how social media data collection and analysis company Geofeedia provided data to police for surveillance purposes, including Twitter data that were used to track recent racially charged protests in Baltimore and Ferguson, Mo.
In a separate matter this past spring, Twitter blocked United States intelligence agencies, including the Central Intelligence Agency, from accessing bulk data via Dataminr. (It’s unclear if the FBI was also involved in that previous deal.)
The social media company told the Guardian at the time: “We have never authorized Dataminr or any third party to sell data to a government or intelligence agency for surveillance purposes.”
Crockford, from the ACLU, said the new FBI-Dataminr deal was curious.
“Just on the heels of Twitter making these courageous decisions to revoke access [in the Geofeedia and CIA cases] we now have the FBI getting this major contract for access through Dataminr,” she said. “It’s confusing, and I really wonder what’s going on at Twitter.”
Crockford said she was worried the new deal may give the FBI access to information about tweets that normally isn’t made visible to all users.
“We’re not totally sure what the FBI is getting access to here,” Crockford said in a phone interview. “We fear they may be getting access to information that Twitter users do not make public, most critically and concerningly: location data.”
Twitter’s firehose includes the contents of tweets, demographic information like gender and interests, the cellular network used, and the users’ geolocation, by latitude and longitude coordinates, if the user tags it, according to a recent Washington Post report.
Crockford said that while monitoring social media may be a powerful tool to combat terrorism and other serious threats, she worries it would be used for other less serious law enforcement investigations, citing the recent examples of protesters and activists being monitored.
“People say dumb stuff on the internet, that doesn’t mean they’re terrorists,” Crockford said.