I can’t wait for the hype around this A.I. bullshit to end. Overblown machine learning without any intelligence whatsoever that increases unemployment and turns the entire internet into corrupted garbage. It wastes insane amounts of electricity, further driving climate disaster. Of course, it also censors anything anyone decides is controversial if you’re a regular user, but if you’re rich enough they’ll sell you a model that has no ethical limitations whatsoever. It’s all just about increasing shareholder value while pouring gasoline on the problems plaguing society. And once this bubble pops, it’ll be some other fraud to raise the value of tech stock. Mark my words, this so-called A.I. will be a colossal disappointment, just like ever other phony innovation peddled by lying capitalists.
I appreciate your honest insight. So much of what we are fighting for is visibility, conversation, and connection. Dehumanizing any group of people systematically puts that at risk. Thankfully, we are inevitable 🩷🤍🩵
My main job is training AI’s via RLHF, we basically assess AI responses to prompts for different AI projects (there isn’t a lot of remote work out there, ya know). One project, I think Meta-based (they don’t always tell us, but sometimes you can figure it out) had a list of Tier 1 and Tier 2 sensitive content. Anything Tier 1 had to be flagged and skipped. Racism, explicit erotica, child abuse, motivation to self-harm or suicide, that kind of thing.
Tier 2 meant “proceed with caution”. One of the examples given for Tier 2 was a fabricated conversation where a gay man comes out to his friends. It wasn’t problematic at all. The AI did a good job. But since it was LGBTQ-related, there was a potential it could get harmful, or erotic, or something else that veered into Tier 1. So even “clean” queer content is treated as if it might turn “risky” at any point. I think this speaks to what you’re talking about.
hm thats really interesting. so in this case, queerness in any context makes the tech companies (and as a result the ai) nervous. thank you for sharing!! 💜
I hope Congress moves as quickly as possible to enact regulations that make AI open, welcoming and safe for members of our LGBTQ+ community and others. In my view, we are witnessing a vacuum because AI is still in its infancy and there's a lack of regulatory oversight. Technology has always been what we make it. AI is no exception. Eventually, I believe we will make it work to serve us.
Thank you for this. It’s freighting when AI is trained like this. That’s especially true for things like content moderation. The AI is blind to queerness and likely won’t recognize slurs against the 2SLGBTQIA+. On the off chance it does recognize a slur, it certainly won’t recognize context of a marginalized person reclaiming it. The answer can’t be pretending 30% of Gen Z doesn’t exist.
yes exactly! i think your connection to content moderation is very astute. the same design dynamics are in play - especially as tech companies move toward trumpian censorship politics from the libertarian viewpoints that they have historically espoused.
I had an AI "art" app on my phone for a while, and the developers kept changing the rules. Something that happened more than once was I found that "lesbian", "gay", "bisexual", "transgender", and "queer" were all banned words.
I can’t wait for the hype around this A.I. bullshit to end. Overblown machine learning without any intelligence whatsoever that increases unemployment and turns the entire internet into corrupted garbage. It wastes insane amounts of electricity, further driving climate disaster. Of course, it also censors anything anyone decides is controversial if you’re a regular user, but if you’re rich enough they’ll sell you a model that has no ethical limitations whatsoever. It’s all just about increasing shareholder value while pouring gasoline on the problems plaguing society. And once this bubble pops, it’ll be some other fraud to raise the value of tech stock. Mark my words, this so-called A.I. will be a colossal disappointment, just like ever other phony innovation peddled by lying capitalists.
I need a name for this feeling of being utterly shocked and not at all surprised at the same time. It’s happening a LOT lately
REJECT CORPORATE FUTURISMS
the only corporate futurism i accept is the costco cinematic universe!
I appreciate your honest insight. So much of what we are fighting for is visibility, conversation, and connection. Dehumanizing any group of people systematically puts that at risk. Thankfully, we are inevitable 🩷🤍🩵
I had no idea. Thank you so much for sharing this.
My main job is training AI’s via RLHF, we basically assess AI responses to prompts for different AI projects (there isn’t a lot of remote work out there, ya know). One project, I think Meta-based (they don’t always tell us, but sometimes you can figure it out) had a list of Tier 1 and Tier 2 sensitive content. Anything Tier 1 had to be flagged and skipped. Racism, explicit erotica, child abuse, motivation to self-harm or suicide, that kind of thing.
Tier 2 meant “proceed with caution”. One of the examples given for Tier 2 was a fabricated conversation where a gay man comes out to his friends. It wasn’t problematic at all. The AI did a good job. But since it was LGBTQ-related, there was a potential it could get harmful, or erotic, or something else that veered into Tier 1. So even “clean” queer content is treated as if it might turn “risky” at any point. I think this speaks to what you’re talking about.
hm thats really interesting. so in this case, queerness in any context makes the tech companies (and as a result the ai) nervous. thank you for sharing!! 💜
I hope Congress moves as quickly as possible to enact regulations that make AI open, welcoming and safe for members of our LGBTQ+ community and others. In my view, we are witnessing a vacuum because AI is still in its infancy and there's a lack of regulatory oversight. Technology has always been what we make it. AI is no exception. Eventually, I believe we will make it work to serve us.
i'm with this
Thank you for this. It’s freighting when AI is trained like this. That’s especially true for things like content moderation. The AI is blind to queerness and likely won’t recognize slurs against the 2SLGBTQIA+. On the off chance it does recognize a slur, it certainly won’t recognize context of a marginalized person reclaiming it. The answer can’t be pretending 30% of Gen Z doesn’t exist.
yes exactly! i think your connection to content moderation is very astute. the same design dynamics are in play - especially as tech companies move toward trumpian censorship politics from the libertarian viewpoints that they have historically espoused.
I had an AI "art" app on my phone for a while, and the developers kept changing the rules. Something that happened more than once was I found that "lesbian", "gay", "bisexual", "transgender", and "queer" were all banned words.
hmmmmm ;;;
My husband’s job was just erased thanks to trump’s agenda.
He’s been researching health disparities within LGBT+ communities for nearly 20 years and all funding is gone.
It’s gone because dear leader doesn’t think we should exist.
be strong and stay safe