Wait, am I Being Fair? Characterizing Deductive Stereotyping and Mitigating It with Fair-GCG
A psychotherapist reported improved relationships with her family after limiting her social media use, while Meta released new smart glasses without Ray-Ban branding, sparking concerns over privacy and surveillance.
A psychotherapist claimed that banning herself from social media significantly improved her relationships with her children and husband. She used to spend 'hundreds of times a day' on her phone[1]. To limit her phone use, she downloaded an app called App Block. Meta, on the other hand, has released new smart glasses without Ray-Ban branding, which can be modded to record short, stealthy clips[2]. The new glasses have raised concerns over privacy and surveillance, with some users on Threads calling them a tool for 'perverts'. However, others argue that the glasses are not capable of 24/7 audio or video surveillance due to battery life limitations. The debate surrounding AI wearables, including smart glasses, pendants, pins, and rings, highlights the tension between their potential benefits and the risks they pose to privacy.
research-papercontroversytool-release
Background sources we checked (1)
- arxiv.org ↗ Warning: This paper contains several toxic and offensive statements. While reasoning generally improves fairness in recent large language models (LLMs), failures persist. In this work, we identify a failure mode, deductive stereotyping, in which models apply population-level stat…