OPINION: The danger is that people, however well-intentioned, take the law into their hands.
By Zwanga Matsila and Colin Thakur
“UNHINGED” and “anarchy” were some terms used to describe the July mayhem in our beloved country. The chaos ostensibly triggered by the jailing of former president Jacob Zuma spiralled into the country’s worst post-democratic era unrest. At least 330 citizens were killed and over 200 shopping malls were devastated.
The overwhelmed police could do little but stand and watch. Citizens and businesses in hitherto unheralded acts of confluence formed vigilante groups to protect people and property. While these vigilante groups had the effect of mitigating some damage and violence, they also fuelled racial discordance. Vigilantism occurs when civilians act in a law enforcement capacity without legal authority.
In the aftermath, South African rekindled the ubuntu spirit creating their own supply chains to freight food hampers of oil, bread and other essentials to citizens without fear or favour. Citizens, spurred by the pandemic-induced cabin fever, also resorted to online crowd-sourcing to find perpetrators.
The process of taking vigilantism online is called digilantism. Facebook and Twitter activists swung into gear, combining their digital skills with vigilantism morphing into digilantes. Digilantism is an area we are researching.
Are there precedents for digilantism?
The most recent was the FBI using digilantism to find and identify the Capitol stormers on January 6 in Washington DC. If they can, so can we, right?
The danger is that people, however well-intentioned, take the law into their hands.
Distributed “evidence” is taken as factual. A likeness of a person is taken to be fait accompli. It hardly seemed to matter that some shared material was recycled from other unrests, from movies, was dated, or even fake. Even worse, when the images are true with respect to date and geo-locations, they were of such poor quality or duration that circumstantial accusations were impossible.
While humans excel at recognising faces, they appear abysmal in recognising photographic faces in photographs which is a mystery to neuroscience and psychology. Even video surveillance is proving problematic as the US government is probing whether surveillance software can be deployed against protesters, with California considering banning the police use of such technology.
Is there any evidence of crime sourcing events through internet images being “wrong”? The 2013 Boston Marathon, with 26 000 participants, was marred by a sadistic bombing killing three runners and wounding hundreds. The marathon inspired digilantism, and our research, as outraged US citizens, tried all means to identify the culprits.
Internet sleuths waged through photo and video-graphic footage and tried to triangulate this with available eyewitness accounts to find the Boston Bombers. The unintended tragic consequences were that they got the identifications horribly wrong, leaving innocent people fearing for their safety, while one person tragically committed suicide.
Charlie Beckett called the Boston fiasco a “media literacy seminar” and hoped that “people are learning to be less stupid. You don’t want to be the person who names a suspect who turns out to be innocent.”
Consider digilantism and the 2018 Moses Mabhida Stadium (MMS) soccer riots. As videos of the MMS rampage went viral and before the police asked for help, internet sleuths began experimenting with reverse image search software such as Google and TinEye to try to identify the hooligans. People felt vindicated as the PSL and police asked for help in identifying some suspects.
The authors themselves tried and did not get many “hits” from the grainy public-posted MMS videos and photographs, although the system reported close comparisons. However, even with the naked eye, the comparisons looked dubious at best. This experiment demonstrated how unreliable the process could be.
The research demonstrates the difficulty of using social photographic and videographic images to apportion accusations affirming the need for a formal academic exercise to be pursued. Does this sound familiar in the current riotous context?
This is the reason unemotional police detective work is crucial. Police have a methodical process to make an accusation, which may well appear pedantic or terribly boring. Further, an allegation must be arbitrated by a court of law. These checks and balances point to a functional democracy.
Social Media does not need the “cry wolf effect”.
The fable belongs to the print era.
*Matsila and Thakur are researching digilantism at the Durban University of Technology. The views expressed are their own.