Social media alerted Spanish police since 2018 of what appeared to be 201 suicide attempts. Direct shots of people, speaking to the camera, with indications that they were going to take their own lives, more or less imminently. A content moderator, conveniently alerted by the tracking bots or by a user who noticed the situation, analyzed the messages that the network user was giving and the context in which he was acting and notified his superiors, who activated the police alert protocols.
Sixty-six of those 201 alerts were real, and 2 actually ended in suicide.
These are the data handled by the inspector of the National Police Corps (CNP) Marta Fernández, who warns of the “zero collaboration” of the platforms – in general – in the prosecution of flagrant crimes.
The notification mechanism is from Interpol, a pan-European collegiate police body, and allows the CNP to act without judicial authorization to act, accessing emails, telephone numbers and IP of the computer from which such content was broadcast.
“When the facts have already been committed, what social networks usually do is eliminate the content because they consider it harmful to their internal policies and relations with users and governments. Of course, it is a system that can be improved,” explains this inspector, consulted by La Vanguardia.
This newspaper revealed last week that around 20% of the Telus staff, the company hired by Meta to control content on Facebook and Instagram, is on sick leave, largely affected by the very harsh content they must filter on the networks. . Suicides, savage revenge between drug gangs, pedophilia, zoophilia...
This newspaper's investigation revealed that, in the same way that content moderators have a protocol to alert their superiors and, if they deem it appropriate, the police authorities, the platforms are highly permissive with what they keep within reach of the public. because it generates traffic.
One of the Telus moderators who was interviewed by La Vanguardia, under anonymity, stated that “Meta's objective is for people to stay as long as possible on the platform, and that is why the policies have been relaxed, they saw that if We eliminated everything, people went to Twitter and lost users. For content, for example sexual exploitation of minors, to be annulled, it must meet some parameters of duration and violence, and to report it to the supervisors, the circumstance must arise that the person committing the crime is the same person who uploaded the video, which it's almost impossible".
But if content presents signs of crime but there is no or appears to be a life at stake, the platforms eliminate it without any alert to the police. “In those cases, the information we receive is very scarce,” Fernández illustrates. “Collaboration can be greatly improved.”
The moderator narrated that labeling a suicide attempt as such so that the police can be alerted “is very difficult because you have to evaluate it live, you have to evaluate whether the person is really going to commit suicide or not, and you don't know because the person He initiates an act only by saying that he does not like life, but there is not enough data for Facebook policies to escalate the case to the police. And in a matter of seconds he can pull out a hidden weapon and blow his head off in front of your eyes and you can't do anything. “Only seeing blood.”
“With the exception of Tik-Tok, which is Chinese, most of the networks are American and are governed by the laws of their country. Things that we would not tolerate here are shielded by freedom of expression,” laments the police inspector.
Another debate that the case has raised is that of the anonymity under which many of those who commit those atrocities act. “From a police point of view, it would be ideal that users would not register with false information; it would be ideal that emails and telephone numbers should be verified,” Fernández proposes.
“These protocols should improve – reflects Fernández – because it would help in prevention, to prevent certain criminals from reoffending or repeating the crime. We are not aware of this being reported to the police. When platforms detect content that violates their policies, they eliminate it, even if there is a possible crime that could be investigated to prevent recurrence or to arrest the person who commits it. Certainly, the networks offer images from the entire planet, so many times it is not our responsibility, they are crimes committed thousands of kilometers away, but of course in some cases it would allow us to intervene."
A recent investigation by The Washington Post into websites that offer open source artificial intelligence demonstrated how easily images or videos of pedophilia can be created. This is the new fear of law enforcement. The speed at which they can be created will lead to a race between artificial intelligences: the one that will create and the one that will filter malicious content, experts predict.
The North American newspaper also included the reflections of Emad Mostaque, executive director of the computer company Stability AI, who appealed to a possibly utopian solution: “Ultimately, it is the responsibility of people as to whether they are ethical, moral and legal in the how they operate this technology.” “The bad things that people will create will be a very, very small percentage of the total use.”
Inspector Fernández is scanning the networks every day and has a pessimistic diagnosis, and not one for the future: “It is alarming. There is more and more violence and, above all, greater tolerance, especially among younger people, who see it on the networks as something natural and normal."
Do you want to share any confidential information with the A Fondo team? Write to us at afondo@lavanguardia.es
