Reproduce “Alignment Myopia in Generative AI” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Remi Denton, Aida Davani, Cynthia L Bennett, Michael Madaio, Roma Patel, Himabindu Lakkaraju
The authors synthesize prior research to argue that conversational AI alignment can be myopic when it optimizes immediate preference signals without accounting for effects of repeated use. They map near-term benefits and potential longer-term risks across cognitive, social, affective, and identity-and-agency facets of well-being. The paper proposes a research agenda, not a validated method or causal test.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.