Papers for

product managers for ai assistants

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Vision language models weigh conflicting text and images with bias

Vision-Language Models Integrate Conflicting Evidence Like Biased Bayesian Observers

Abstract: Vision-language models must reconcile conflicting text and images. Prior work has characterized the resulting behavior in terms of modality biases--e.g., a "blind faith in text"--but why preferences vary across models, domains, and input conditions remains unclear. We develop a predictive account from a Bayesian perspective, positing that VLMs combine conditionally independent evidence from each modality with prior beliefs, subject to modality-specific biases. The resulting two-parameter cue combination model accurately predicts VLM preferences under unseen combinations of input degradation and transfers to new forms of unreliability without refitting. Separating evidence strength from integration bias explains how VLMs can follow text even while weighting visual evidence more heavily, reconciling reports of text and vision dominance under a single combination rule.

Wed 14 OctComputation and LanguageArtificial IntelligenceComputer Vision and Pattern Recognition
The gist
When vision-language models see images and read text, they sometimes get conflicting information. The authors explain this by comparing the models to biased Bayesian observers that combine clues from both sources but with their own preferences. Their two-parameter model can predict how these models choose between images and text even when the quality of one changes. This helps clarify why sometimes the model trusts text more and other times it relies on visual cues.
Open → 2610.19407v1

Users and AI share reasoning tasks with varying involvement preferences

What Did the AI Take On? Characterizing Cognitive Delegation in LLM Reasoning

Abstract: Large language models (LLMs) often perform intermediate cognitive work while carrying out users' requests, yet it remains unclear which parts users intended to delegate and how they wanted to remain involved. This matters because consequential choices may go unnoticed, limiting users' ability to steer the process, while reviewing every step would make delegation burdensome. We examined this with 24 LLM users across three knowledge-work tasks, collecting 992 retrospective annotations of reasoning steps. From this, we developed taxonomies of LLM cognitive work, delegation enactment, and desired delegation protocols at the reasoning-step level. Our analysis revealed that participants viewed about half of all steps (48.6%) as AI-initiated, meaning the AI took on work they had not requested. Desired involvement varied with cognitive work and delegation enactment, even when contributions matched participants' intent. We propose design implications and sketches for supporting more deliberate cognitive delegation through flexible protocols and inspectable, revisable AI-initiated decisions.

Mon 5 OctHuman-Computer Interaction
The gist
When people use large language models (LLMs) like AI helpers, the AI often does extra thinking steps that users didn’t directly ask for. The authors studied how users feel about which thinking tasks the AI should do and how much they want to be involved. They found that users saw almost half of the AI’s actions as taking initiative without being told. People want different levels of participation depending on the type of work and how the AI took charge. The authors suggest ways to design AI systems so users can better control and review what the AI does.
Open → 2610.06328v1