Illustration of the US Social Network Instagram logo on a tablet screen.
Kirill Kudryavtsev | AFP | Getty images
Meta Excused Thursday and said he had corrected an “error” that had led some Instagram users to report a flow of violent and graphic content recommended on their personal “rolled” page.
“We have corrected an error that led some users to see content in their Instagram roller flow which should not have been recommended. We apologize for the error,” said a Meta spokesperson in a press release shared with CNBC.
The declaration comes after a number of Instagram users went to various social media platforms to express concerns about a recent influx of violent and “not sure for work” in their flows.
Some users have claimed to have seen such content, even with the “sensitive content control” of Instagram activated at its highest moderation parameter.
According to Meta-politicalThe company strives to protect users from disturbing images and delete particularly violent or graphic content.
The prohibited content includes videos “representing dismemberments, visible bowels or charred bodies”, as well as content which contains “sadistic remarks to imagery representing the suffering of humans and animals”.
However, Meta says that this allows a certain graphic content if it helps users to condemn and raise awareness of important questions such as human rights violations, armed conflicts or acts of terrorism. Such content can be accompanied by limitations, such as warning labels.
On Wednesday evening in the United States, CNBC was able to see several messages on Instagram coils which seemed to show corpses, graphic injuries and violent attacks. The publications were labeled “sensitive content”.
According to the Meta website, he uses internal technology and a team Of more than 15,000 examiners to help detect worrying images, technology contributing to publications.
Technology, which includes artificial intelligence and automatic learning toolsMay detect and detect proactively “the vast majority of the content violation” even before users report it, the website indicates.
In addition, Meta works to avoid Recommend content on its platforms This can be “of low quality, reprehensible, sensitive or inappropriate for young viewers,” he said.
Exchange policy
The error with Instagram’s content recommendation, however, comes after Meta said that she changed the way he has policies on his platforms to promote freedom of expression.
In a statement Published on January 7, the company said that it changed the way it had applied its content policies to reduce errors that had led to censorship on its platforms.
Meta said that he changed his automated digitization systems “for all violations of policies” to focus on “illegal and high severity violations, such as terrorism, sexual exploitation of children, drugs, fraud and fraud”. For less serious political violations, the company said it would be based on users to report problems before acting.
Meanwhile, Meta said that her systems drew too much content according to the predictions that she “could” violate the standards and that she was “getting rid of most of these demotions”.
CEO Mark Zuckerberg also announced that the company would authorize more political content and would change its program to verify third -party facts with a “community notes” model, similar to the system that exists on Elon MuskPlatform X.
The movements were largely considered as an effort to Zuckerberg Tie-liens With American President Donald Trump, who criticized Meta’s moderation policies in the past.
According to a Meta spokesperson on xThe CEO visited the White House earlier this month “to discuss the way META can help administration defend and advance American technological leaders abroad”.
As part of a wave of technological layoffs in 2022 and 2023, Meta Cut 21,000 employeesalmost a quarter of his workforce, which affected a large part of his Civic integrity and confidence and safety teams.
