Leaked training documents obtained by The Intercept offer more details about Meta’s updated Hateful Conduct rules. Specific examples of speech Meta allows include “Immigrants are grubby, filthy pieces of shit,” “Jews are flat out greedier than Christians,” and “Trans people are immoral.”
404 Media has updates from unnamed employees on how the moderation changes have been received:
“The entire thread of comments shared is dissent toward the new policy, save for one leader repeating Zuckerberg talking points. I’d call the mood shock and disbelief,” they added. “It’s embarrassment and shame that feels self-inflicted, different than mistakes the company has made in the past.”
Oh, and as for Joel Kaplan’s More Speech and Fewer Mistakes, Meta is already answering an “error” that blanked out search results for “LGBT” and “Trans.”
After user outrage, Meta is deleting its small army of AI bot profiles in order to fix a bug that removed the ability to block the accounts. As I reported earlier today, the bots have been around since 2023, but many users just discovered them in the last few days following news coverage and social media posts.
The New York Times reports on a spate of cases where predators posing as photographers and promoters used social media to connect with parents and solicit child sexual abuse imagery, taking advantage of not just Instagram but also parents who want to turn their young children into influencers.
[The New York Times]
