How to Avoid Debate: Scalable AI Safety via Doubly-Efficient Interactive Proofs

Liyan Chen, Yael Tauman Kalai, Zoe Xi
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:14831-14858, 2026.

Abstract

As AI models continue to develop powerful capabilities, it becomes critical that we are able to verify that their output is aligned with our intentions. A recent line of work focuses on verification via debate, a model of interactive proofs where two competing powerful provers, or AI models, debate each other to convince a weak verifier, or a human, of the correctness of their claim. However, debate assumes that the two AI models possess equal abilities and that one of them is truthful, which may not be realistic. In this work, we show how to avoid debate: we initiate the study of single-prover interactive proofs for AI safety. Prior results in single-prover interactive proofs do not immediately carry over to the AI safety setting because they do not work when the computation has access to an oracle, such as to human judgment or an external database such as the web. We present doubly-efficient single-prover interactive proofs for oracle-aided computations (also known as relativizing proofs), in the settings where (1) the computation is robust, in the sense that the output does not change if at most a small fraction of the answers to oracle queries are incorrect, or (2) the oracle is a low-degree polynomial. These results suggest that interactive verification is possible even without debate, under structured or noise-tolerant oracle access.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chen26bc, title = {How to Avoid Debate: Scalable {AI} Safety via Doubly-Efficient Interactive Proofs}, author = {Chen, Liyan and Kalai, Yael Tauman and Xi, Zoe}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {14831--14858}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chen26bc/chen26bc.pdf}, url = {https://proceedings.mlr.press/v306/chen26bc.html}, abstract = {As AI models continue to develop powerful capabilities, it becomes critical that we are able to verify that their output is aligned with our intentions. A recent line of work focuses on verification via debate, a model of interactive proofs where two competing powerful provers, or AI models, debate each other to convince a weak verifier, or a human, of the correctness of their claim. However, debate assumes that the two AI models possess equal abilities and that one of them is truthful, which may not be realistic. In this work, we show how to avoid debate: we initiate the study of single-prover interactive proofs for AI safety. Prior results in single-prover interactive proofs do not immediately carry over to the AI safety setting because they do not work when the computation has access to an oracle, such as to human judgment or an external database such as the web. We present doubly-efficient single-prover interactive proofs for oracle-aided computations (also known as relativizing proofs), in the settings where (1) the computation is robust, in the sense that the output does not change if at most a small fraction of the answers to oracle queries are incorrect, or (2) the oracle is a low-degree polynomial. These results suggest that interactive verification is possible even without debate, under structured or noise-tolerant oracle access.} }
Endnote
%0 Conference Paper %T How to Avoid Debate: Scalable AI Safety via Doubly-Efficient Interactive Proofs %A Liyan Chen %A Yael Tauman Kalai %A Zoe Xi %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chen26bc %I PMLR %P 14831--14858 %U https://proceedings.mlr.press/v306/chen26bc.html %V 306 %X As AI models continue to develop powerful capabilities, it becomes critical that we are able to verify that their output is aligned with our intentions. A recent line of work focuses on verification via debate, a model of interactive proofs where two competing powerful provers, or AI models, debate each other to convince a weak verifier, or a human, of the correctness of their claim. However, debate assumes that the two AI models possess equal abilities and that one of them is truthful, which may not be realistic. In this work, we show how to avoid debate: we initiate the study of single-prover interactive proofs for AI safety. Prior results in single-prover interactive proofs do not immediately carry over to the AI safety setting because they do not work when the computation has access to an oracle, such as to human judgment or an external database such as the web. We present doubly-efficient single-prover interactive proofs for oracle-aided computations (also known as relativizing proofs), in the settings where (1) the computation is robust, in the sense that the output does not change if at most a small fraction of the answers to oracle queries are incorrect, or (2) the oracle is a low-degree polynomial. These results suggest that interactive verification is possible even without debate, under structured or noise-tolerant oracle access.
APA
Chen, L., Kalai, Y.T. & Xi, Z.. (2026). How to Avoid Debate: Scalable AI Safety via Doubly-Efficient Interactive Proofs. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:14831-14858 Available from https://proceedings.mlr.press/v306/chen26bc.html.

Related Material