Tag
The article discusses the absence of outcome-based, user-facing safety labels for AI, similar to food labels, and suggests requirements for such certification, including use-case specificity, human rights as a non-negotiable floor, and independent auditing.
Binance highlights its support efforts, noting it has recovered $8.2B across 1.28M user appeals for funds sent to wrong addresses since 2021, with CZ emphasizing user protection.
This paper introduces AISPA, a user-centric framework for auditing system prompts in commercial LLM applications. The audit of 3,249 instructions across 88 products reveals inconsistent protective coverage, shallow adoption, and pervasive problematic instructions.