and the longest continuous internet shutdown ever recorded & moderating genocidal content targeting your people. One moderator saw his cousin being murdered in a video routed to him. He wasn’t even allowed to take a short mental health break.
CULTURE
-
Google’s Content Moderation Gap in Under-resourced Languages
By
–
I told people @ Google about this & asked them to build language technology & have human moderation, in so-called "under-resourced languages” like Tigrinya. Although we were at Google, concerns sent to YouTube went into a void never to be acknowledged.
-
DAIR Research on Tigray Genocide and AI Ethics Concerns
By
–
Announcing new work from DAIR which is very close to my heart, 3 years in the making. When #TigrayGenocide, the deadliest genocide of the 21st century thus far, started in November 2020, it was 1 month before I got fired from Google.
-
Data-First Culture: Human-in-the-Loop Approach to Unlock Client Potential
By
–
A data-first culture should maximize efficiency, using time & expertise to unlock clients' data potential. 🤝
— Snorkel AI (@SnorkelAI) 15 avril 2025
Snorkel AI Co-Founder @henryehrenberg shared insights at #HumanX2025 on fostering this culture with a human-in-the-loop approach. Full video: https://t.co/ddfa4V37Mr pic.twitter.com/9mBlI2IGKIA data-first culture should maximize efficiency, using time & expertise to unlock clients' data potential. Snorkel AI Co-Founder @henryehrenberg shared insights at #HumanX2025 on fostering this culture with a human-in-the-loop approach. Full video: https://
bit.ly/3RFnb2L -
AI transforms celebrities into Roman statues with stunning realism
By
–
I turned every celebrity into a Roman statue using AI. And the results? Insanely realistic. Here’s how you can do it too: 1. Michael Jackson
-
AI Mirrors Society’s Biases: LLMs Struggle to Keep Pace
By
–
AI acts as a mirror. It reflects our society's pre-conceived notions and biases right back at us. Societies keep evolving and LLMs will not be able to keep pace. So, here’s what we need to do…
-
AI Image Generation Model Perpetuates Indian Stereotype Bias
By
–
So, I asked the model to add diversity of age, gender, and race, including Indians, it produced Image 2 (below). The model added Indian faces with red ‘bindis’ on their foreheads as if it were mandatory for Indian executive attire! Clearly we have a problem.
-
Addressing AI Image Generation Bias in Diverse Demographics
By
–
The bad news is AI has biases. The good news is we can fix them. I asked a specific model to generate an image of senior Asian executives sitting around a round table. I won't name the model, as the idea is to focus on the issue rather than criticize any company. The model
-
Why AI Speech Recognition Got Worse Despite Hype
By
–
In Scientific American, @asmelashteka & I ask: "In the age of unbridled AI hype, with the likes of Elon Musk claiming to build a “maximally truth-seeking AI,” how did we come to have less reliable speech recognition systems than we did before?"
-
Cool Research Links Earth Intelligence to Hitchhiker’s Guide Concept
By
–
This is really cool work! But also it's confirming that Hitchhiker's Guide to the Galaxy may've accurately described the nature of intelligence here on Earth