The FBI is seeking to develop an early-warning system based on material “scraped” from social networks.
It says the application should provide information about possible domestic and global threats superimposed onto maps “using mash-up technology”.
The bureau has asked contractors to suggest possible solutions including the estimated cost.
Privacy campaigners say they are concerned that the move could have implications for free speech.
The FBI’s Strategic Information and Operations Center (SOIC) posted its “Social Media Application” market research request onto the web on 19 January, and it was subsequently flagged up by New Scientist magazine.
The document says: “Social media has become a primary source of intelligence because it has become the premier first response to key events and the primal alert to possible developing situations.”
It says the application should collect “open source” information and have the ability to:
- Provide an automated search and scrape capability of social networks including Facebook and Twitter.
- Allow users to create new keyword searches.
- Display different levels of threats as alerts on maps, possibly using colour coding to distinguish priority. Google Maps 3D and Yahoo Maps are listed among the “preferred” mapping options.
- Plot a wide range of domestic and global terror data.
- Immediately translate foreign language tweets into English.
The FBI says the information would be used to help it to predict the likely actions of “bad actors”, detect instances of people deliberately misleading law enforcement officers and spot the vulnerabilities of suspect groups.
The FBI issued the request three weeks after the US Department of Homeland Security released a separate report into the privacy implications of monitoring social media websites.
It justified the principle of using information that users have provided and not opted to make private.
“Information posted to social media websites is publicly accessible and voluntarily generated. Thus the opportunity not to provide information exists prior to the informational post by the user,” it says.
It noted that the department’s National Operations Center had a policy in place to edit out any gathered information which fell outside of the categories relevant to its investigations.
It listed websites that the centre planned to monitor. They include YouTube, the photo service Flickr, and Itstrending.com – a site which shows popular shared items on Facebook.
It also highlighted words it looked out for. These include “gangs”, “small pox”, “leak”, “recall” and “2600” – an apparent reference to the hacking-focused magazine.
The London-based campaign group, Privacy International, said it was worried about the consequences of such activities.
“Social networks are about connecting people with other people – if one person is the target of police monitoring, there will be a dragnet effect in which dozens, even hundreds, of innocent users also come under surveillance,” said Gus Hosein, the group’s executive director.
“It is not necessarily the case that the more information law enforcement officers have, the safer we will be.
“Police may well find themselves overwhelmed by a flood of personal information, information that is precious to those it concerns but useless for the purposes of crime prevention.”
The group noted that it was seeking information from the UK’s Metropolitan Police Service about its use of social networks.