Despite evidence that state and federal law enforcement agencies have contracted with data analytics firms to mine information from social media users, policymakers have still not moved to protect Americans’ constitutional rights to freedom of speech and association from the misuse of “Big Data,” according to an expert on information and privacy law.
The revelations underline the need for new approaches to transparency in order to ensure that police use of Big Data remains within “acceptable boundaries,” Teresa Scassa, Canada Research Chair in Information Law and Policy at the University of Ottawa, wrote in a paper posted in Scripted. The disclosures from the American Civil Liberties Union (ACLU) and the Brennan Center for Justice in 2016 did not say what kind of information police are collecting, or how they use it.
According to Scassa, the questions that need answering before tackling police use of data mining services include:
- Is metadata culled from private social media companies like Twitter and Facebook still public once it is sold to an company like Geofeedia?
- When the algorithms and software used by private companies to process the data constitute trade secrets, is the end product they sell to clients truly “publicly available” or “publicly accessible?”
- Even when individual pieces of information are public, does the full picture afforded by large data sets trespass on privacy?
Private Industry Self-Policing
Scassa went on to note that while we’re still chasing down the shifting boundaries of public and private information, we lack a system of governance in place to hold industry and law enforcement accountable for what kind of data they purchase, and how they use it.
When the ACLU raised an alarm over law enforcement contracts with the data analytics company Geofeedia, which offered geographic data on private citizens culled from social media accounts like Twitter and Facebook, the backlash was immediate. Social media companies scrambled to revise their policies, wanting to distance themselves as much as possible from police surveillance activities.
Facebook, Instagram, and Twitter all terminated Geofeedia’s access to public user data, and a week later, Geofeedia announced it had laid off half of its workforce.
Other third party data mining companies that were hauled into the spotlight, including Snaptrands and Dataminr, also pivoted in terms of how they marketed their services.
While “Snaptrends’ website has not referenced any relationships with law enforcement or national security,” since March of 2017, writes Scassa, the company still identifies “public safety organizations” as a sector that can benefit from its data.
Facebook’s Platform Policy for developers now stipulates that developers should not “use data obtained from us to provide tools that are used for surveillance.” But compliance is voluntary, lacking any legislative teeth or transparency mechanisms. While Facebook can audit 3rd party usage of its data, the public cannot.
“The public nature of social media data undermines traditional oversight and transparency paradigms, particularly those based on privacy norms,” writes Scassa. “This is in large part because social media data is considered to be public.”
The use of this data is relatively low-cost compared to other types of surveillance activities, and for the time being, sidesteps the burdensome warrant process. Until analytics companies came on the scene, law enforcement could perform manual searches of social media without a warrant, but other kinds of searches might require a paper trail, allowing some accountability in the courts.
Analytics applied to large volumes of data “can be quite sophisticated and can include mapping, predictive policing, profiling, the use of facial recognition software, and so on,” writes Scassa. And what sets it apart from other types of information garnered by law enforcement is its potential use for predictive policing, as opposed to being focused on a specific event– raising “issues of targeting and profiling.”
“Public disclosures of police service contracts do not reveal” how public data is processed or used; and emails obtained by the UCLA to and from 63 police departments across the country in the Geofeedia case only show how the company marketed its services. In emails, Geofeedia cited protests following the deaths of Freddie Grey and Michael Brown “as examples of its usefulness to police services,” a fact that Scassa says “aggravated a climate of mistrust, racial division, and a sense that the authorities were using surveillance and profiling to target minority communities.”
In addition to privacy concerns, “big data analytics used in surveillance may involve constitutional values such as the freedoms of speech and association, as well as anti-discrimination values… for example, a search algorithm that includes the hashtag “BlackLivesMatter” identifies individuals on the basis of both political speech and association with a group or movement.”
The full article, Law Enforcement in the Age of Big Data and Surveillance Intermediaries: Transparency Challenges, is available for free download here. This summary was prepared by Deputy Editor Victoria Mckenzie. She welcomes readers’ comments.