Search papers, labs, and topics across Lattice.
This paper critiques the reliance on proxy utility functions in decision theory, highlighting specific conditions where such optimization can lead to detrimental outcomes. The author argues that proxy optimization can misalign incentives and produce suboptimal decisions, ultimately undermining the effectiveness of decision-making frameworks. Key findings suggest that understanding the limitations of proxy functions is crucial for developing more robust decision-theoretic models.
Relying on proxy utility functions can lead to harmful decision-making outcomes that undermine the very goals they aim to achieve.
I discuss conditions under which maximizing a proxy utility function is harmful and suggest this poses problems for applying decision theory.
