AI erases evidence of human rights abuses on social media

A shocking investigation by the BBC has revealed that evidence of potential human rights abuses is being permanently erased after being erased by technology companies’ artificial intelligence (AI) tools.

While the platforms say they can retain artwork when it’s in the public interest, the BBC says it has had a mixed experience trying to post footage documenting attacks on civilians in Ukraine. The footage was quickly deleted despite its value in exposing the reality of events.

AI has been effective in removing harmful and illegal content on a large scale. However, when it comes to moderating the violent images of wars, machines fail to identify the nuances involving human rights violations.

Human rights groups say it is urgent that social media companies prevent this vital information from disappearing. “It’s understandable that they develop machines to delete difficult or traumatic content,” says Rusbridger.

Meta’s Supervisory Board, created by Mark Zuckerberg and known as the independent “supreme court” for the company, is asked how to develop machines, whether human or artificial intelligence, to make more reasonable decisions.

US Ambassador for Global Criminal Justice Beth Van Schaak points out that tech companies aren’t denied the right to monitor their content, but concern arises when that information suddenly goes missing.

A former travel journalist, Ihor Zakharenko, witnessed this problem in Ukraine. After the Russian invasion, he documented attacks on civilians. However, when he tried to post the videos on social media, they were promptly removed.

To avoid losing this crucial evidence, human rights organizations such as Mnemonic have developed tools to automatically download and save evidence of human rights violations on social media. However, they cannot cover all conflict areas in the world.

Activists say a formal system needs to be established to securely collect and store deleted content, including metadata to verify the authenticity of evidence. It is essential that social media platforms establish agreements with accountability mechanisms around the world in order to preserve this information for possible future legal actions.

With information from BBC.

The post AI deletes evidence of human rights violations on social networks first appeared on Olhar Digital.

Source: Olhar Digital