Saved in:
Bibliographic Details
Main Authors: Brown-Cohen, Jonah, Irving, Geoffrey, Marshall, Simon C., Newman, Ilan, Piliouras, Georgios, Szegedy, Mario
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.08630
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910016589529088
author Brown-Cohen, Jonah
Irving, Geoffrey
Marshall, Simon C.
Newman, Ilan
Piliouras, Georgios
Szegedy, Mario
author_facet Brown-Cohen, Jonah
Irving, Geoffrey
Marshall, Simon C.
Newman, Ilan
Piliouras, Georgios
Szegedy, Mario
contents AI safety via debate uses two competing models to help a human judge verify complex computational tasks. Previous work has established what problems debate can solve in principle, but has not analysed the practical cost of human oversight: how many queries must the judge make to the debate transcript? We introduce Debate Query Complexity}(DQC), the minimum number of bits a verifier must inspect to correctly decide a debate. Surprisingly, we find that PSPACE/poly (the class of problems which debate can efficiently decide) is precisely the class of functions decidable with O(log n) queries. This characterisation shows that debate is remarkably query-efficient: even for highly complex problems, logarithmic oversight suffices. We also establish that functions depending on all their input bits require Omega(log n) queries, and that any function computable by a circuit of size s satisfies DQC(f) <= log(s) + 3. Interestingly, this last result implies that proving DQC lower bounds of log(n) + 6 for languages in P would yield new circuit lower bounds, connecting debate query complexity to central questions in circuit complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08630
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Debate is efficient with your time
Brown-Cohen, Jonah
Irving, Geoffrey
Marshall, Simon C.
Newman, Ilan
Piliouras, Georgios
Szegedy, Mario
Artificial Intelligence
Computational Complexity
AI safety via debate uses two competing models to help a human judge verify complex computational tasks. Previous work has established what problems debate can solve in principle, but has not analysed the practical cost of human oversight: how many queries must the judge make to the debate transcript? We introduce Debate Query Complexity}(DQC), the minimum number of bits a verifier must inspect to correctly decide a debate. Surprisingly, we find that PSPACE/poly (the class of problems which debate can efficiently decide) is precisely the class of functions decidable with O(log n) queries. This characterisation shows that debate is remarkably query-efficient: even for highly complex problems, logarithmic oversight suffices. We also establish that functions depending on all their input bits require Omega(log n) queries, and that any function computable by a circuit of size s satisfies DQC(f) <= log(s) + 3. Interestingly, this last result implies that proving DQC lower bounds of log(n) + 6 for languages in P would yield new circuit lower bounds, connecting debate query complexity to central questions in circuit complexity.
title Debate is efficient with your time
topic Artificial Intelligence
Computational Complexity
url https://arxiv.org/abs/2602.08630