👉 "To maintain its credibility as a leader in responsible #AI, #Google needs to justify this change in policy and outline in detail which safeguards it will adopt to ensure its compliance with #humanrights and humanitarian principles." @marianaolaizola.bsky.social bhr.stern.nyu.edu/quick-take/g...

Happy to read your longer piece. I definitely appreciate hearing different perspectives but it’s easier if we know which sentences and sources we are talking about. The thread makes it hard for sure

Inside Meta’s dehumanizing new speech policies for trans people "A trans person isn't a he or she, it's an it," reads a new guideline telling moderators what is now allowed on Facebook and Instagram

You’re right - I should have linked to this article by Casey Newton which specifies the types of statements that Meta now considers non-violating, some of which fall into the category of hate speech (e.g., "Trans people are freaks.") www.platformer.news/meta-new-tra...

You cited the report by the special rapporteur on freedom of expression, so I cited the same... It seems to me you just want to “win” an argument, not engage in good faith debate.

No. neither of those examples is hate speech. See my paragraph under “The need for nuance.”
Also, fwiw, I don’t say Meta is allowing hate speech - I say Meta is no longer committed to enforcing against it.

Happy to consider your views. I value this exchange

Finally, when Meta says “we remove serious insults…” etc they mean reactive removals, right? Which takes us back to the argument that relying on reactive moderation alone is not sufficient

The “Mexican” example is a perfect instance of why companies can’t rely on AI alone. A human reviewer would have corrected that error immediately (if companies had more of them)

Second, why do you assume that proactive detection means using AI exclusively? The rapporteur noted the danger of increasing reliance on AI detection alone, but this doesn’t mean that a combination of AI plus human review is not compatible with IHRL and the rapporteur’s comments specifically

Thanks for your views, Pranesh. A few questions: Isn’t it the case that “lower-severity violations” encompassed hate speech? Zuckerberg mentioned as high-severity only posts dealing with drugs, terrorism, CSAM. So that part of the argument holds, in my view. Second, …

You’re right, Pranesh. But what percentage of hate speech gets reported? Meta has previously bragged about how effective its automated tools were at detecting violations. And now it’s simply rolling them back and shifting the burden on users. Happy to consider your perspective, though.

My latest for Tech Policy Press: How Meta Turned its Back on Human Rights

How Meta Turned Its Back on Human Rights | TechPolicy.Press Mark Zuckerberg has chosen to disregard longstanding human rights norms in favor of a simplistic approach to free speech, writes Mariana Olaizola Rosenblat.

Mark Zuckerberg has chosen to disregard longstanding human rights norms in favor of a simplistic approach to free speech, writes NYU Stern Center for Business and Human Rights policy advisor on technology and law Mariana Olaizola Rosenblat. It's not too late to change course.

Some Facts About Fact-Checking: Defending the Imperfect Search for Truth in an Era of Institutionalized Lying | TechPolicy.Press Fact-checking ensures a distinction between facts and falsehoods and that everything is not up for grabs because powerful people say so, writes Paul Barrett.

Fact-checking signals that not everything is up for grabs because powerful people say so, writes NYU Stern Center for Business and Human Rights deputy director Paul Barrett. "Meta has scorned and stifled that signal. Others should not follow suit."

Is Bluesky set up to deal with misinformation? As the number of Bluesky users skyrockets, how will it handle misinformation and content moderation?

Asked for my thoughts on the Bluesky boomlet, I told Logically Facts that this platform is doing well, but as it grows, it will need to staff up on content moderation and fine tune its filtering tech. Learn from the mistakes of Facebook, Brand X, et al.: www.logicallyfacts.com/en/article/b...

I'm pleased to join a panel discussion tonight (Nov. 19) at 6pm at the University of Maryland's Smith School of Business on "Social Media Business Models & the Promotion of Extremism." Please join us in College Park! www.rhsmith.umd.edu/events/socia...