April 16, 2024

This Week in Responsible AI: Apr 16, 2024

This Week in Responsible AI: Apr 16, 2024

Law/Policy

• English translation of China's "Basic Safety Requirements for Generative Artificial Intelligence Services"

• The China Academy of Information and Communications Technology's results from its AI safety benchmark

• ‘British DARPA’ to build AI gatekeepers for ‘quantitative safety guarantees’

• Public Policy at Hugging Face

• Texas is replacing thousands of human exam graders with AI

• AI was supposed to make police bodycams better. What happened?

Datasets

• The Nine Lives of ImageNet: A Sociotechnical Retrospective of a Foundation Dataset and the Limits of Automated Essentialism

• The AI Village Defcon red-teaming dataset has been released

• Adobe’s ‘Ethical’ Firefly AI Was Trained on Midjourney Images

Explainability

• Does Transformer Interpretability Transfer to RNNs?

• LeGrad: An Explainability Method for Vision Transformers via Feature Formation Sensitivity

Bias

• A biased test kept thousands of Black people from getting a kidney transplant. It’s finally changing

• I’m still trying to generate an AI Asian man and white woman. Why is it so difficult?

Other

• new first monday issue: Ideologies of AI and the consolidation of power

• "We demonstrate that adding a strategic text sequence (STS)---a carefully crafted message---to a product's information page can significantly increase its likelihood of being listed as the LLM's top recommendation"

• To understand the risks posed by AI, follow the money

Compiled by Leif Hancox-Li