Click to learn what we’re doing to help organizations respond to COVID-19.

As previously published on our blog...

Privacy, Civil Liberties, and Video Analytics: Part 1

During the past year and a half, Palantir has been developing Palantir Video, a video analytics application that will allow Palantir Gotham users to better integrate video information into the rest of their analyses. The initial Palantir Video functionality will be limited to basic playback capabilities and the ability for a user to tag individuals and events within video data and associate those tags with information from other data sources in Palantir Gotham. As we continue to develop this capability, we will explore more advanced analytic capabilities, many of which may raise concerns about the potential effects of this kind of analysis on the protection of privacy and civil liberties.

Recognizing this, we convened some of our privacy and civil liberties advisors—a group of academics and advocates who we consult on new product developments that was recently formalized as the Palantir Council of Advisors on Privacy and Civil Liberties—to discuss video analytics. Over the course of several posts, we will present the issues raised by the group concerning video analytics in general and Palantir Video in particular, as well as how we might be able to use Palantir’s capabilities to support policies governing the use of video data that better protect privacy and civil liberties.1

In this post we will explore, at a very basic level, the global growth of video surveillance use, the current law on the use of video surveillance, and the privacy and civil liberties concerns that can accompany the deployment of video surveillance capabilities.

The Growth of Video Surveillance

The use of video surveillance has grown substantially in the last decade. In the United States, more and more major metropolitan areas have created CCTV systems. Authorities typically create these systems by both installing new cameras and tapping into existing surveillance networks operated by transit authorities, schools, and even private companies.2 In the United Kingdom, which began using CCTV in the 1990s, a 2011 report estimated that there are more than 1.85 million cameras operating in the country—one for every 32 U.K. citizens. This includes not just law enforcement run cameras but also private sector systems.3 A recent market research report predicted that the global CCTV market will reach around $23.5 billion by the end of 2014, which, the report suggests, represents a compounded annual growth rate of 20.5%.4

The rapid growth of CCTV systems is fueled by their perceived utility in supporting effective law enforcement efforts. Law enforcement officials argue that the use of video surveillance can deter crime, improve officer response time, and facilitate the after-the-fact investigation of crimes.5 They frequently cite anecdotal evidence of particular crimes that have been prevented or solved with the help of CCTV.6 Studies of CCTV effectiveness are sometimes less enthusiastic, generally finding “a modest but significant desirable effect on crime”7 or at best suggesting that the impact of CCTV has been “variable.”8</a>

The Current State of Law Governing Video Surveillance

When trained on public spaces, CCTV surveillance systems are minimally constrained in their use. Indeed, one privacy expert concludes, “Meaningful legal strictures on government use of public surveillance cameras in Great Britain, Canada, and the United States are nonexistent.”9 In the United States, the question of the legality of public surveillance was effectively settled in United States vs. Knotts, in which the Supreme Court considered the legality of tracking a car’s movements with an electronic beeper. In its Opinion, the Court concluded that “a person traveling… on public thoroughfares has no reasonable expectation of privacy in his movements from one place to another.”10 The private sector use of surveillance cameras in public spaces is similarly unconstrained, with the only real restriction coming from so-called “peeping Tom” statutes that criminalize filming or recording of certain activities where a person has some expectation of privacy.11 The International Association of Chiefs of Police has developed guidelines for the use of CCTV for public safety and community policing, and some jurisdictions have adopted these or similar guidelines.12 While useful, these guidelines are not legally binding.

The recent Supreme Court decision in United States v. Jones may indicate a potential avenue for judicial constraint of ubiquitous CCTV monitoring. Some commentators have suggested that a reading of the various opinions in that case suggests a majority of support for the “mosaic theory” of privacy. The theory argues that while the surveillance of individual acts taking place in public may not be constitutionally protected, “the collective sum of the different acts over time amount to a search” and therefore such surveillance should be subject to the Fourth Amendment.13 As CCTV becomes more and more prevalent, one could envision a scenario in which the use of cameras to track an individual over an extended period of time is analyzed under the mosaic theory. However, any judicial adoption of this theory will have to await the next relevant case, which could be years in coming.

Video Surveillance and Threats to Privacy and Civil Liberties

The use of video surveillance systems by government agencies in a public space raises a number of privacy and civil liberties concerns such as:

  • Generally speaking, surveillance can have a negative societal effect by inducing individuals to conform to societal norms and pushing individual choices towards the “bland and the mainstream.”14
  • Even when in a public space, there is a “state of privacy” that flows from a degree of anonymity.15 In other words, people behave differently when they are among strangers than they might when they are with people who recognize them. CCTV monitoring, combined with video analytics that facilitate identification, could diminish an individual’s sense of anonymity thereby inhibiting his or her freedom of action in a public space.
  • CCTV surveillance is generally indiscriminate. It is “often akin to a dragnet search, which can ensnare a significant amount of data beyond that which was originally sought.”16
  • Given this indiscriminate nature, there will be a temptation for substantial mission creep in which cameras deployed for one reason (e.g., to prevent and solve crime) become tasked with other missions that may be substantially different from this original intent. Such function creep is often not subject to the same public scrutiny as the initial camera installation. This is not a new phenomenon in surveillance. As observed by the Church Committee four decades ago, “We have seen a consistent pattern in which programs initiated with limited goals, such as preventing criminal violence or identifying foreign spies, were expanded to what witnesses characterized as ‘vacuum cleaners,’ sweeping in information about lawful activities of American citizens.”17
  • As with any surveillance tool, CCTV can be abused by bad actors within an organization. Individuals with access to CCTV data may use the tools at their disposal to steal information, stalk an ex-spouse, or otherwise abuse their power. Any surveillance system must be designed with an awareness of this risk, however uncommon these abuses may be in actual practice.

With these issues in mind, we turned to our advisors for help. How do video analytics contribute to these P/CL concerns? Are there other concerns raised by the development and deployment of sophisticated data analytics? How could we design Palantir Video to help mitigate these concerns?

We learned a great deal from our discussion and will share some of that with you in our next few posts.

  1. Not all members of the Palantir Council of Advisors on Privacy and Civil Liberties attended this discussion. Participants are free to discuss the meeting publicly if they choose, provided that they do not disclose any proprietary or otherwise confidential information (no information was so designated in this discussion). In order to encourage an open and frank exchange, Palantir will not publicly attribute any comments to individuals or organizations participating in the meetings. 

  2. Alex Johnson, “Smile! More and more, you’re on camera,” MSNBC, June 25, 2008.

  3. “Big Brother is DEFINITELY watching you,” Daily Mail, March 3, 2011. Others estimate the number of cameras to be as high as 4.2 million. See, John Woodhouse, CCTV and its effectiveness in tackling crime, House of Commons Library, July 1, 2010.

  4. Nicholas Bombourg, “Global CCTV Market Forecast to 2014,” PRNewswire, March 13, 2012.

  5. Department of Homeland Security, CCTV Developing Privacy Best Practices, December 17/18, 2007.

  6. See e.g., Allison Klein, “Police Go Live Monitoring D.C. Crime Cameras,” Washington Post, February 11, 2008.

  7. Welsh, Brandon C. and David P. Farrington, Effects of Closed Circuit Television Surveillance on Crime, Campbell SystematicReviews, 2008. 

  8. Martin Gill and Angela Spriggs, Assessing the impact of CCTV, Home Office Report, February 2005. 

  9. Christopher Slobogin, Privacy at Risk (Chicago: University of Chicago Press, 2007) 89. 

  10. United States v. Knotts, 460 U.S. 276, 281 (1983). 

  11. Slobogin 89. 

  12.;tabid=423. See also, NYPD Domain Awareness System Public Security Privacy Guidelines (, UK CCTV Code of Practice ( 

  13. Orin Kerr, “What’s the Status of the Mosaic Theory After Jones?” The Volokh Conspiracy, January 23, 2012.

  14. Julie Cohen, Examined Lives: Informational Privacy and the Subject as Object, 52 Stanford L. Rev. 1373, 1426 (2000). 

  15. Allan Westin, Privacy and Freedom, (New York: Atheneum, 1967) 31. 

  16. Dan Solove, Nothing to Hide, (New Haven: Yale UP, 2011) 179. 

  17. Select Committee to Study Governmental Operations with respect to Intelligence Activities, “Intelligence Activities and the Rights of Americans,” U.S. Senate, April 26, 1976, pp 3 – 4.