Hunting scammers on social networks generates debate about the use of technology; understand

by

The Bolsonarist acts marked by the depredation of the buildings of the three Powers in Brasília aroused an investigative impetus on social networks, with names and photos of extreme right-wing radicals in advertisements by authorities, reports, publications of influencers and anonymous profiles.

Experts heard by Sheet claim that volunteer work to gather evidence about the January 8 attack can help authorities, but also involve innocent people and that it is only the state’s responsibility to identify criminals.

On the 9th, the Ministry of Justice released an official channel for complaints about the destruction of the buildings of Congress, the Palácio do Planalto and the STF (Federal Supreme Court), which can be registered at the email [email protected]. The agency says it received 50,000 messages on the first day of operation.

On Instagram alone, publications with the terms “terrorist”, “coup plotters”, “radicals” and “extreme right” received 78.5 million interactions in the week between January 8 and 14, according to CrowdTangle, Meta’s platform that compiles data about the social network.

This wave of interest has fueled the growth of pages on Twitter and Instagram. The main one of these profiles, Contragolpe Brasil, brings photos and data of people who would be involved in the act and recommends registering complaints on the official government channel. Created on Sunday, the profile reached 891,000 followers on the first day and reached 1.2 million followers on Monday (16).

Instagram blocked last Tuesday (11) the publication of images and texts on the timeline, but the restriction fell on Friday (13).

When contacted by the report, Meta said it would not comment on specific cases and profiles. The social network’s community guidelines do not cite the identification of allegedly involved in crime, it only asks that its users comply with the law.

The Contragolpe Brasil profile did not respond to the questions asked by the Sheet on Instagram, the only channel available for contact.

What can be disclosed?

The disclosure of photos and names of possible people involved in Sunday’s attacks does not constitute a criminal practice, when one considers that the vandals themselves released images of themselves, according to InternetLab director Heloísa Massaro. The public interest further expands the possibilities of use.

She says profiles involved in independent investigations have not been shown to be profit-oriented. Therefore, they do not need to comply with the LGPD (General Law for the Protection of Personal Data).

Massaro recalls that public safety matters are outside the scope of the LGPD. A criminal data protection law is under discussion in Congress, which will deal with criminal offenses.

Identifying participants in coup acts without proper verification can harm innocent people. That would be framed as falsely charging, a crime, punishable by up to eight years.

Founder of Novelo Data, Guilherme Felitti recalls the aftermath of the 2013 Boston Marathon attack. The New York Post tabloid published a photo of then high school student Salah Barhoum, circulated as a suspect in the terrorist act. The young man of Moroccan descent has suffered threats and persecution.

According to the American vehicle, the image had been forwarded between the police forces and indicated two possible suspects. An investigation by BuzzFeed News later showed that allegations against Barhoum had started on Reddit and 4chan forums. The New York Post spoke at the time: it maintained its version of the story and stated that it did not identify the young people, but removed the word suspects from the title.

What to worry about when producing or consuming this content?

Data science researcher at the Oxford Internet Institute Adriano Belisário states that transparency is fundamental to disclosing sensitive information, such as the identity of those involved in alleged crimes.

The profiles have published photos and names of those allegedly involved in the acts without explaining how they achieved such a result. the posts started on Sunday, before the government released the first list of those involved, on Tuesday (10).

Felitti says that facial recognition tools can also be applied to the images to identify those involved, despite warning that these computer models deliver wrong answers at a dangerous frequency.

And whoever operates these algorithms is unable to avoid errors, since the artificial intelligence receives the information and delivers the most likely answer — there is always an inaccuracy, greater for black people.

What can be done with the materials available on the networks?

The authorities themselves have used images of the invasion published by Bolsonarists to gather evidence about the coup act.

The federal intervenor in Public Security of the Federal District, Ricardo Capelli, said in an interview with journalists this Monday that the videos serve as evidence to advance the investigations of the attacks on the 8th.

Belisário claims that publications on social networks provide other clues: where and when the scammers were. This information can be used to reconstruct the crime scene in news reports and documentary productions.

That’s what the New York Times did in the documentary Days of Rage (Days of Rage, in free translation) about the attack on the Capitol to prevent the inauguration of Democrat Joe Biden. The US newspaper’s visual investigations team developed animations based on times and images from live broadcasts.

The key actors in the invasion of the Capitol were identified from the official investigation. The US federal police claims to have analyzed material equivalent to 361 uninterrupted days of recordings, which add up to 9 terabytes of information.

You May Also Like

Recommended for you