Pith. sign in

REVIEW 2 cited by

On Goodhart's law, with an application to value alignment

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.09638 v1 pith:TMOI5QCM submitted 2024-10-12 stat.ML cs.AIcs.LGmath.STstat.TH

classification stat.MLcs.AIcs.LGmath.STstat.TH
keywords goodhartmeasuregoalpoliciestruewhenassessdistinction
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

``When a measure becomes a target, it ceases to be a good measure'', this adage is known as {\it Goodhart's law}. In this paper, we investigate formally this law and prove that it critically depends on the tail distribution of the discrepancy between the true goal and the measure that is optimized. Discrepancies with long-tail distributions favor a Goodhart's law, that is, the optimization of the measure can have a counter-productive effect on the goal. We provide a formal setting to assess Goodhart's law by studying the asymptotic behavior of the correlation between the goal and the measure, as the measure is optimized. Moreover, we introduce a distinction between a {\it weak} Goodhart's law, when over-optimizing the metric is useless for the true goal, and a {\it strong} Goodhart's law, when over-optimizing the metric is harmful for the true goal. A distinction which we prove to depend on the tail distribution. We stress the implications of this result to large-scale decision making and policies that are (and have to be) based on metrics, and propose numerous research directions to better assess the safety of such policies in general, and to the particularly concerning case where these policies are automated with algorithms.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Toward a Theory of Value in AI Alignment

    cs.AI 2026-08 conditional novelty 6.0 of 10

    A systematic annotation of 94 AI alignment papers shows the field largely equates human values with measurable preferences, rarely defines values, and is increasingly removing humans from alignment evaluation.

  2. A Case for Specialisation in Non-Human Entities

    cs.CY 2025-02 conditional novelty 6.0 of 10

    A position paper making the case that specialised, well-specified AI systems are more robust, secure, and governable than general-purpose AGI systems, and that hard-to-specify tasks need specified governance.

Pith tools