ArticleFrontiers in artificial intelligence2026
Large-scale conversational data auditability via cryptographic commitments and hash-chained audit-ledger verification.
Article in Frontiers in artificial intelligence, 2026. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Not yet cited in PubMed.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
0 citing papers in PubMed.
No citing paper in PubMed yet.
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
4 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
The auditability of large language model-based systems requires verifiable evidence without exposing the sensitive content of interactions. This article proposes a hash-chained audit-ledger framework to transform public LLM conversations into canonical off-chain units, generate cryptographic commitments using SHA-256 and HMAC-SHA-256, build a lightweight ledger with Merkle roots, and verify integrity against induced tampering. The study integrates three heterogeneous sources, LMSYS-Chat-1M, WildChat, and Chatbot Arena, and produces 3,971,887 auditable units. In the full-corpus HMAC-SHA-256 experiment, the 11,899.55 MiB normalized repository was represented by 1,473.44 MiB of hash-only audit records, 2.03 MiB of ledger block headers, and 380.52 MiB for the operational ledger plus membership structures required for record-level verification. Five induced tampering categories, Merkle and decision-hash integrity, hash-chain linkage integrity, block ordering integrity, membership completeness, and dataset-membership consistency, were detected without false acceptance; a separate clean-ledger baseline validated the untampered ledger with zero error signals. Load experiments showed near-linear growth of mean total pipeline time with respect to the number of audit records. Operational ledger-stage throughput increased across the intermediate workloads but declined at the largest evaluated load, while full record-set verification throughput remained comparatively stable up to 250,000 interactions before decreasing in the 500,000-interaction scenario. This contribution demonstrates that the traceability of decisions generated by LLM can be formulated as a measurable, private, and cryptographically verifiable experimental chain. The framework provides a documented numerical-audit basis for independent consistency assessment, regulatory compliance, and forensic analysis when ledger states are evaluated against preserved artifacts, trusted checkpoints, or externally anchored commitments.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.