Abstract
Individuals performing open source investigations can misidentify alleged perpetrators and dox innocent parties online, which can subsequently lead to threats and harassment against innocent parties and their loved ones. For example, threats were made against Sunil Tripathi’s family after he was wrongly identified as one of the Boston Marathon bombers and doxed on Reddit and Twitter. In 2020, the Berkeley Protocol on Digital Open Source Investigations was published as a guide, and it includes a set of principles to govern open source investigations. However, the Berkeley Protocol is limited to open source investigations performed by those working for organizations. It does not include situations in which organizations invite volunteers to help, also known as crowdsourcing investigations, or when individuals are performing their own investigations without any organizational support, such as when individuals Tweet people’s pictures and accuse them of crimes while asking others to help with identifying these alleged perpetrators. This gap leaves innocent parties at risk of the consequences of misidentification and doxing committed on the internet.
This article examines some of the best practices of three different organizations (Europol, Bellingcat, and Trace Labs) in their use of crowdsourcing in open source investigations and then considers how to adapt those best practices to better respect the Berkeley Protocol principles of accountability, competency, and accuracy. This article also looks at whether it is possible to adapt the Berkeley Protocol to better protect innocent parties from individuals performing investigations outside the scope of an organization. It finds that the Berkeley Protocol can and should be expanded to better guide organizations that instigate crowdsourcing, but that eliminating the risks of misidentification and doxing committed by individuals working outside the scope of an organization would be nearly impossible.
Recommended Citation
Leigh M. Dannhauser,
Protecting the Innocent: How to Prevent the Consequences of Misidentification and Doxing by Volunteers Helping with Open Source Investigations,
32
Cath. U. J. L. & Tech
1
(2024).
Available at:
https://scholarship.law.edu/jlt/vol32/iss2/3
Included in
Civil Rights and Discrimination Commons, Communications Law Commons, Computer Law Commons, Criminal Law Commons, Entertainment, Arts, and Sports Law Commons, First Amendment Commons, Fourth Amendment Commons, Internet Law Commons, Law Enforcement and Corrections Commons, National Security Law Commons, Privacy Law Commons, Science and Technology Law Commons