December 13, 2023

This Week in Responsible AI: Dec 13, 2023

This Week in Responsible AI: Dec 13, 2023

If any readers happen to also be at NeurIPS, happy to say hi :-).

Policy

• Artificial intelligence act: Council and Parliament strike a deal on the first rules for AI in the world

• MIT group releases white papers on governance of AI

• Considerations for Governing Open Foundation Models

• Mind the Gap: The Challenges of Assurance for Artificial Intelligence

• U.S. Government Accountability Office provides recommendations on how government agencies should acquire and use AI systems.

• UK: Judges Given the OK to Use ChatGPT in Legal Rulings

Privacy

• Hashmarks: Privacy-Preserving Benchmarks for High-Stakes AI Evaluation

• Partnership on AI report: How to Analyze Data while Preserving User Privacy

• AI and Mass Spying

AI Harms

• France’s Digital Inquisition

• AI-Generated ‘Undressing’ Images Move from Niche Pornography Discussion Forums to a Scaled and Monetized Online Business

• "This notebook contains code to estimate the CO2 emissions associated with air travel to ACL 2023. We estimate that ACL's 2500 attendees flew 23.6 million kilometers, amounting to approximately 2400 metric tons of CO2 emissions."

Authenticity

• Why virtual isn’t actual, especially when it comes to friends

• The New Turing Test

Fairness

• On The Fairness Impacts of Hardware Selection in Machine Learning

• Entity-Based Evaluation of Political Bias in Automatic Summarization

Transparency

• Google's Gemini continues the dangerous obfuscation of AI technology

• "We find on average that audiences perceive news labeled as AI-generated as less trustworthy, not more, even when articles themselves are not evaluated as any less accurate or unfair."

• Announcing Purple Llama: Towards open trust and safety in the new world of generative AI

Other

• Assessing Prompt Injection Risks in 200+ Custom GPTs

• "if an AWS employee were to answer my questions the way Amazon Q does, I imagine Andy Jassy would parachute into the room to personally fire them on the spot"

Compiled by Leif Hancox-Li