May 25, 2023

This Week in Responsible AI: May 25, 2023

This Week in Responsible AI: May 25, 2023

New Research

• Humans and algorithms work together---so study them together

• A review of the many ways to estimate Shapley values

Policy

• What can the EU learn from China's generative AI regulation before it adopts its AI Act?

• Want to Know About the Future of AI Equity? Watch What’s Happening in California

Diversity

• Why AI’s diversity crisis matters, and how to tackle it

• After OpenAI hearing, A.I. experts urge Congress to listen to more diverse voices on regulation. See also: How Congress Fell for Sam Altman’s AI Magic Tricks, The Senate’s hearing on AI regulation was dangerously friendly

• When Does the Gender Gap Start in the Computer Science Field?

Privacy

• FTC Says Ed Tech Provider Edmodo Unlawfully Used Children’s Personal Information for Advertising and Outsourced Compliance to School Districts

• Facebook to be fined £648m for mishandling user information

• FTC Warns About Misuses of Biometric Information and Harm to Consumers

AI Harms

• Remember when Google mislabelled black people as gorillas? 8 years on, you still can't search for gorillas in photo apps.

• AI threatens humanity’s future, 61% of Americans say: Reuters/Ipsos poll

• The presuppositions underlying OpenAI's governance document

• Debt Collectors Want To Use AI Chatbots To Hustle People For Money

Generative AI

• The Electronic Privacy Information Center releases a report [PDF] on the harms of generative AI

• He Helped Train ChatGPT. It Traumatized Him.

• The Fanfic Sex Trope That Caught a Plundering AI Red-Handed

• How con artists use AI, apps, social engineering to target parents, grandparents for theft

• Alien Minds, Immaculate Bullshit, Outstanding Questions

• Introduction to red teaming large language models (LLMs)

• "50% of generative search engine responses lack supportive citations, and 25% of the citations provided are off point"

• Situating current arguments on AI 'plagiarism' in the historical evolution of concepts of authorship and originality

• From Theory to Reality: Explaining the Best Prompt Injection Proof of Concept

• Lost in Translation: Large Language Models in Non-English Content Analysis

Compiled by Leif Hancox-Li