August 20, 2023

This Week in Responsible AI: Aug 20, 2023

This Week in Responsible AI: Aug 20, 2023

Law/Policy

• The New York Times prohibits using its content to train AI models

• Ad firm plans to use people’s data in a maneuver to sink data privacy bill

• FTC comments on digital ownership and generative AI

• CFPB to develop rules on data brokers

• Where's the liability in harmful AI speech?

• China’s new AI regulations begin to take effect

• The AI Power Paradox: Can States Learn to Govern Artificial Intelligence—Before It’s Too Late?

Privacy

• After Backlash, Zoom Now Says It Won’t Use Any Customer Content to Train AI Systems

• "One former Worldcoin employee attributed its difficulty in scanning Asian eyes to the lack of diverse training data, noting that the system was trained mostly on the eyes of white and Black people."

• Amazon’s ‘Just Walk Out’ Tech Is a Privacy Nightmare

Is AI bad science?

• Timnit Gebru Looks at Corporate AI and Sees a Lot of Bad Science

• Can emotion recognition models be accurate?

• Mind your Language (Model): Fact-Checking LLMs and their Role in NLP Research and Practice

Open-source = ethical?

• The mirage of open-source AI: Analyzing META's LLAMA 2 release strategy

• Open (For Business): Big Tech, Concentrated Power, and the Political Economy of Open AI

Bias and Harms

• Algorithm overlooks Spanish speakers in online SNAP ads

• AI Detection Tools Falsely Accuse International Students of Cheating

• AI Causes Real Harm. Let’s Focus on That over the End-of-Humanity Hype

Ethics

• Inside the messy ethics of making war with machines

• AI Systems Must Not Confuse Users about Their Sentience or Moral Status

• The Artificiality of Alignment

Accountability

• Dislocated accountabilities in the “AI supply chain”: Modularity and developers’ notions of responsibility

• Can We Red Team Our Way to AI Accountability?. See also: What happens when thousands of hackers try to break AI chatbots

Labor

• Clickworkers are turning against each other

• Unleashing possibilities, ignoring risks: Why we need tools to manage AI’s impact on jobs

Generative AI

• An urgent care clinic in New York is responding to declining patient volumes by flooding the internet with AI-generated posts on fictional topics like the "Derek Jeter Herpes Tree"

• Paper exams, chatbot bans: Colleges seek to ‘ChatGPT-proof’ assignments

• Does ChatGPT have a liberal bias?

• The Imminent Enshittification of the Internet

• Automatic Disinformation Threatens Democracy— and It’s Here

Compiled by Leif Hancox-Li