When researchers searched for the 1980s skinhead band Evil Skins on YouTube Music, it surfaced the group’s song “Zyklon Army.” The title is a reference to Zyklon B, the cyanide-based gas used in Germany’s extermination camps during World War II, and the tune opens with the Nazi salute “Sieg Heil” chanted over and over. The song was illustrated with a video auto-generated by YouTube, which has software that creates videos to go along with songs that don’t have videos of their own.
But first, YouTube Music showed an ad for McDonald’s, touting the fast-food chain’s frozen Fanta drinks.
That’s just one of the more colorful details in a report published Wednesday by a pair of nonprofits, the Anti-Defamation League and the Tech Transparency Project. Building on an August report that examined how social media sites recommend hateful and white supremacist content to users, the new paper focuses on sites that serve paid advertisements alongside that content, potentially profiting from it.
The report finds that Google-owned YouTube is the worst offender among four major social platforms when it comes to running ads on content associated with hateful extremism and white supremacy. Elon Musk’s X came in second, while Meta’s Facebook and Instagram were the least likely to show ads on such material.
On each platform, the researchers plugged in 130 search terms they’ve identified as the names of hate groups and movements to see what came up. They found that YouTube served ads on the search results page for 67 of those terms, including some ads from well-known brands. Instagram and Facebook displayed ads on only a few of them, while X was in between.
Tech companies often argue it’s impossible to police every user-generated post that violates their policies. But Yael Eisenstat, vice president of the ADL’s Center for Technology and Society, said Meta’s relative success at keeping ads off white supremacist content “shows that it is absolutely possible to fix some of these problems.”
In a previous report, the researchers created personas on each social platform to see which ones would recommend hateful content via their algorithms. In that case, it was YouTube that performed best, while Meta’s platforms lagged behind. That dovetails with recent research that found YouTube may have largely solved its infamous “rabbit hole” problem, in which its algorithms had a penchant for recommending increasingly extreme content to users who started off viewing relatively innocuous videos.
The report comes as Elon Musk has been blaming the ADL for a sharp drop in advertising revenue on X since he bought it almost a year ago.
Advertising industry insiders told my colleague Elizabeth Dwoskin recently that the real culprit in scaring off advertisers is Musk’s own statements and policies, which have emboldened bigots and restored previously banned accounts. Eisenstat said the ADL’s goal with these reports is not to punish social platforms but to bring to light problems so the platforms can address them.
“I doubt that any of these companies actually want to have ads run alongside violative content because it’s not good for business,” Eisenstat said. “But that is different from actually prioritizing and resourcing the teams and tools necessary to ensure that’s not happening.”
Last week, Israeli Prime Minister Benjamin Netanyahu urged Musk to denounce antisemitism in a live-streamed meeting at a Tesla factory in Fremont, Calif. (Musk deflected.) And earlier this week, a group of some 100 Jewish leaders published a letter calling out Musk for allegedly abetting antisemitism on his platform. Elad Nehorai, a Jewish writer and activist who organized the letter, told The Tech 202 that Musk has not responded so far.
In a statement, Google spokesperson Davis Thompson said, “We don’t allow ads to run on our platforms that include or promote hateful content, and we have strict policies prohibiting hate speech on YouTube. We take this issue seriously and have terminated several YouTube channels noted in ADL’s report. This review is ongoing, and we may take further action to remove content from our platform.”
A request for comment from X returned an auto-response that said, “Busy now, please check back later.”
source: https://www.msn.com/en-us/news/poli...deos-for-white-supremacist-groups/ar-AA1howTA
archive: https://archive.md/wip/vMUoM
But first, YouTube Music showed an ad for McDonald’s, touting the fast-food chain’s frozen Fanta drinks.
That’s just one of the more colorful details in a report published Wednesday by a pair of nonprofits, the Anti-Defamation League and the Tech Transparency Project. Building on an August report that examined how social media sites recommend hateful and white supremacist content to users, the new paper focuses on sites that serve paid advertisements alongside that content, potentially profiting from it.
The report finds that Google-owned YouTube is the worst offender among four major social platforms when it comes to running ads on content associated with hateful extremism and white supremacy. Elon Musk’s X came in second, while Meta’s Facebook and Instagram were the least likely to show ads on such material.
On each platform, the researchers plugged in 130 search terms they’ve identified as the names of hate groups and movements to see what came up. They found that YouTube served ads on the search results page for 67 of those terms, including some ads from well-known brands. Instagram and Facebook displayed ads on only a few of them, while X was in between.
Tech companies often argue it’s impossible to police every user-generated post that violates their policies. But Yael Eisenstat, vice president of the ADL’s Center for Technology and Society, said Meta’s relative success at keeping ads off white supremacist content “shows that it is absolutely possible to fix some of these problems.”
In a previous report, the researchers created personas on each social platform to see which ones would recommend hateful content via their algorithms. In that case, it was YouTube that performed best, while Meta’s platforms lagged behind. That dovetails with recent research that found YouTube may have largely solved its infamous “rabbit hole” problem, in which its algorithms had a penchant for recommending increasingly extreme content to users who started off viewing relatively innocuous videos.
The report comes as Elon Musk has been blaming the ADL for a sharp drop in advertising revenue on X since he bought it almost a year ago.
Advertising industry insiders told my colleague Elizabeth Dwoskin recently that the real culprit in scaring off advertisers is Musk’s own statements and policies, which have emboldened bigots and restored previously banned accounts. Eisenstat said the ADL’s goal with these reports is not to punish social platforms but to bring to light problems so the platforms can address them.
“I doubt that any of these companies actually want to have ads run alongside violative content because it’s not good for business,” Eisenstat said. “But that is different from actually prioritizing and resourcing the teams and tools necessary to ensure that’s not happening.”
Last week, Israeli Prime Minister Benjamin Netanyahu urged Musk to denounce antisemitism in a live-streamed meeting at a Tesla factory in Fremont, Calif. (Musk deflected.) And earlier this week, a group of some 100 Jewish leaders published a letter calling out Musk for allegedly abetting antisemitism on his platform. Elad Nehorai, a Jewish writer and activist who organized the letter, told The Tech 202 that Musk has not responded so far.
In a statement, Google spokesperson Davis Thompson said, “We don’t allow ads to run on our platforms that include or promote hateful content, and we have strict policies prohibiting hate speech on YouTube. We take this issue seriously and have terminated several YouTube channels noted in ADL’s report. This review is ongoing, and we may take further action to remove content from our platform.”
A request for comment from X returned an auto-response that said, “Busy now, please check back later.”
source: https://www.msn.com/en-us/news/poli...deos-for-white-supremacist-groups/ar-AA1howTA
archive: https://archive.md/wip/vMUoM
