Audio-language models (ALMs) are increasingly used as judges for speech-to-speech systems, but a judge that receives audio may not actually use paralinguistic evidence. We introduce counterfactual audits for paralinguistic response evaluation.
Large Audio-Language Models (LALMs) have been widely used as judge models for the automatic evaluation of generated speech. However, prior approaches predominantly focus on holistic naturalness, leaving fine-grained paralinguistic distinctions underexplored.
The paper introduces a validation‑gated audit framework for voice AI customer‑care systems, treating them as stateful, multi‑turn, tool‑mediated interactions where bias and safety can manifest as added burdens before a final decision. The framework distinguishes between native speech‑to‑speech, cascaded ASR‑to‑LM‑to‑TTS, and hybrid architectures, and applies matched service facts across controlled caller presentation conditions to validate fact invariance, presentation cues, artifacts, and acoustic measurements. It outlines seven validation gates, a six‑family metric set, and demonstrates the approach with a synthetic refund‑dispute audit example, while noting that production results are withheld until the protocol is satisfied.
arXiv:2609.23416v1 Announce Type: cross Abstract: Long-form audio performance is often summarized by context length and aggregate accuracy, obscuring how language, evidence, and task jointly shape di...
arXiv:2608. 19515v1 Announce Type: new Abstract: Prosodic cues can convey task-relevant information that alters the trajectory and outcome of a task-oriented dialogue, even when the words themselves remain unchanged.
arXiv:2609.08899v2 Announce Type: replace-cross Abstract: Speech deepfakes can mimic a speaker's voice convincingly enough to deceive listeners and automated systems. This has driven strong progress...
Prosodic cues can convey task-relevant information that alters the trajectory and outcome of a task-oriented dialogue, even when the words themselves remain unchanged. Yet existing benchmarks typically evaluate prosodic perception, response appropriateness, and task-oriented dialogue in isolation, making it difficult to test whether prosodic evidence changes downstream decisions.
The paper introduces Inquesto Score (IS), a protocol that measures voice‑agent reliability by calculating the percentage of calls that reach the caller’s goal without functional failure. IS defines explicit failure events and severity levels, evaluates timing, semantic, and state‑dependent failures using audio, scenario predicates, tool traces, and a pinned open‑model judge, and provides diagnostic views on behavior, acoustic robustness, identity handling, and speaker groups. The authors evaluate IS v0.1 on 30 scenarios, three acoustic conditions, four speaker groups, and 306 calls per agent across 13 configurations, demonstrating that reliable measurement requires evidence beyond transcripts and explicit treatment of deployment conditions.
The paper evaluates the use of large language models (LLMs) as judges for assessing conversational voice agents, comparing human judgments with GPT‑4.1 and GPT‑5 across telecom and retail interactions. It examines agreement, metric‑level correlations, and consistency across three evaluation configurations (p0, p1, p2) to determine how reliably LLMs can judge conversational quality and safety. The study finds that LLM‑based evaluation can be effective but its reliability varies by metric and configuration, suggesting a hybrid approach where LLMs handle scalable assessment while humans focus on metrics requiring contextual interpretation.
arXiv:2608. 09930v1 Announce Type: cross Abstract: Automated Text-to-Speech (TTS) evaluation methods (Mean Opinion Score (MOS) predictors and Audio Large Language Models (Audio-LLM) judges) are expected to reflect human perception, yet it is unclear how well they capture the distinct aspects of speech that listeners actually perceive.
The paper introduces VeriSpeak, a benchmark of 3,879 spoken claims for evaluating fact verification in Large Audio Language Models (LALMs). It shows a clear modality gap: models that verify written claims well often fail on spoken versions, and retrieval alone offers limited improvement. Combining retrieval with explicit reasoning yields the best performance, reaching 86.1% accuracy and demonstrating the need for grounded reasoning over retrieved evidence in speech misinformation detection.
arXiv:2608. 06409v1 Announce Type: cross Abstract: Speech language models are increasingly evaluated on paralinguistic tasks by the accuracy of prompted answers, but answer accuracy combines failures at different stages of the audio-to-answer computation.
The paper introduces ECHO, a paired diagnostic benchmark for evaluating Chinese real‑time spoken dialogue systems on turn‑taking. ECHO pairs examples that share the same overlap transcript but differ in preceding multi‑turn context, requiring either Yield or Keep actions, and also includes off‑talk cases to test unnecessary yielding. Experiments on four speech systems reveal that three systems over‑yield, correctly keeping the floor on fewer than 13% of backchannels, while the fourth system shows a more balanced performance, illustrating that interruption‑only evaluation can overestimate turn‑taking reliability.