[討論] How do federal agencies share information collected from social media

密西

Federal agen cies may share inform a tion they collect from social media across all levels of govern ment and the private sector and will some times even disclose data to foreign govern ments (for instance, iden ti fi ers on travel and immig ration forms). In partic u lar, inform a tion is shared domest ic ally with state and local law enforce ment, includ ing through fusion centers, which are post -9/11 surveil lance and intel ligence hubs that were inten ded to facil it ate coordin a tion among federal, state, and local law enforce ment and private industry. Such unfettered data shar ing magni fies the risks of abus ive prac tices.

Part of the risk stems from the dissem in a tion of data to actors with a docu mented history of discrim in at ory surveil lance, such as fusion centers. A 2012 bipar tisan Senate invest ig a tion concluded that fusion centers have “yiel ded little, if any, bene fit to federal coun terter ror ism intel ligence efforts, ” instead produ cing reams of low-quality inform a tion while labeling Muslim Amer ic ans enga ging in innoc u ous activ it ies, such as voter regis tration, as poten tial threats. More recently, fusion centers have been caught monit or ing racial and social justice organizers and protests and promot ing fake social media posts by right-wing provocateurs as cred ible intel li gence regard ing poten tial viol ence at anti-police brutal ity protests. Further, many police depart ments that get inform a tion from social media through fusion centers (or from federal agen cies like the FBI and DHS directly) have a history of targeting and surveilling minor ity communit ie s and activists, but lack basic policies that govern their use of social media. Finally, exist ing agree ments permit the U.S. govern ment to share social media data — collected from U.S. visa applic ants, for example — with repress ive foreign govern ments that are known to retali ate against online crit ics.

The broad dissem in a tion of social media data ampli fies some of the harms of social media monit or ing by elim in at ing context and safe guards. Under some circum stances, a govern ment offi cial who initially reviews and collects inform a tion from social media may better under stand — from witness inter views, notes of obser vations from the field, or other mater ial obtained during an invest ig a tion, for example — its mean ing and relev ance than a down stream recip i ent lack ing this back ground. And any safe guards the initial agency places upon its monit or ing and collection — use and retention limit a tions, data security proto cols, etc. — cannot be guar an teed after it dissem in ates what has been gathered. Once social media is dissem in ated, the origin at ing agency has little control over how such inform ation is used, how long it is kept, whether it could be misin ter preted, or how it might spur over reach.

Together, these dynam ics amplify the harms to free express sion and privacy that social media monit or ing gener ates. A qual i fied and poten tially unre li able assess ment based on social media that a protest could turn viol ent or that a partic u lar person poses a threat might easily turn into a justific ation for poli cing that protest aggressively or arrest ing the person, as illus trated by the examples above. Simil arly, a person who has applied for a U.S. visa or been invest ig ated by federal author it ies, even if they are cleared, is likely to be wary of what they say on social media well into the future if they know that there is no endpoint to poten tial scru tiny or disclos ure of their online activity. Formerly, one branch of DHS I&A had a prac tice of redacting publicly avail able U.S. person inform a tion contained in open-source intel ligence reports dissem in ated to part ners because of the “risk of civil rights and liber ties issues.” This prac tice was an appar ent justi fic a tion for remov ing pre-public a tion over sight to identify such issues, which implies that DHS recog nized that inform a tion identifi fy ing a person could be used to target them without a legit im ate law enforce ment reason.