August 10, 2023

This Week in Responsible AI: Aug 10, 2023

This Week in Responsible AI: Aug 10, 2023

I was on vacation, so today's update is a belated one covering the last ~3 weeks!

Research

• Are Model Explanations Useful in Practice? Rethinking How to Support Human-ML Interactions.

• A Human Rights-Based Approach to Responsible AI

• The shaky foundations of large language models and foundation models for electronic health records

• On the transparency of large AI models

• White Paper: AI Outputs and the First Amendment

• When Personalization Harms Performance: Reconsidering the Use of Group Attributes in Prediction

Law/Policy

• New deal on EU-US data flows sparks privacy fears and business uncertainty

• Generative AI services pulled from Apple App Store in China ahead of new regulations

• Equal Employment Opportunity Commission (EEOC) enters into a first of its kind consent decree with a tutoring company for algorithmic/AI age discrimination

• Seven AI companies commit to safeguards at the White House's request

• The AI rules that US policymakers are considering, explained

• UK’s approach to AI safety lacks credibility, report warns

• TikTok’s algorithm will be optional in Europe

• The movement to limit face recognition tech might finally get a win

• AI regulation is taking shape, but startups are being left out

• Zero Trust AI Governance

• The AI Crackdown is Coming

• A.I. Microdirectives Could Soon Be Used for Law Enforcement

• Why Chinese entities are turning to People’s Daily censorship AI to avoid political mines

AI Harms

• Q&A: When automation in government services fails

• In Mannheim, an automated system reports hugs to the police

• Why watermarking AI-generated content won’t guarantee trust online

• Sneak Preview: A blueprint for an AI Bill of Rights for Education

• Unregulated AI Will Worsen Inequality, Warns Nobel-Winning Economist Joseph Stiglitz

• Does social media polarize voters? Unprecedented experiments on Facebook users reveal surprises

• Six ways that AI could change politics

• We need a Weizenbaum test for AI

• Don't Let the Math Distract You: Together, We Can Fight Algorithmic Injustice

• Justine Bateman on AI, Labor, and the Future of Entertainment

• Automated Firing & Algorithmic Management: Mounting a Resistance, with Veena Dubal, Zephyr Teachout and Zubin Soleimany | AI Now Salons

Privacy

• How to Quickly Get to the Important Truth Inside Any Privacy Policy

• DHS Used Clearview AI Facial Recognition In Thousands Of Child Exploitation Cold Cases

• Convicted fraudster Martin Shkreli is touting a medical AI chatbot—much to experts’ concern

• Climate Justice and Labor Rights | Part I: AI Supply Chains and Workflows

Generative AI

• The human decisions that shape generative AI: Who is accountable for what?

• The tricky truth about how generative AI uses your data

• Remini tops the App Store for its viral ‘AI headshots’ but its body edits go too far, some say

• Authors are losing their patience with AI, part 349235. See also: The Fear Of AI Just Killed A Very Useful Tool

• AI language models are rife with political biases

• Researchers Identify False Twitter Personas Likely Powered by ChatGPT

• Language Is a Poor Heuristic For Intelligence

• Exploring Generative AI

Compiled by Leif Hancox-Li