Human-AI Collaboration

Language models are increasingly being deployed as user simulators, but their memory is far more reliable than that of real users. To measure this gap, we run a series of classic memory experiments from psychology on both humans and language models. Across tasks, we find that out-of-the-box language models exhibit better memory than humans, even when prompted to imitate human behavior. We then show that better prompting strategies and the use of a compactor can cause language models to forget content in a more human-like way. Using these methods, we show preliminary evidence that language models with human-like memory constraints can function as more effective user simulators in a downstream education task. Finally, we release human reference data and benchmarks to support future work on simulating human memory with language models.

Current AI interaction models treat the prompt as the primary object of exchange, omitting a critical layer: the user's latent source intent, the goal state preceding and motivating the prompt. Here we introduce Intent Signal Theory (IST), a computational framework that formalises this missing intent layer. IST distinguishes four objects routinely conflated: latent source intent (I*), observable intent proxy (I-hat), encoded carrier (P), and model output (O). It formalises dimensional weights, encoding masks, structural and fidelity recovery scores, and public-private intent decomposition. The Theorem of Irreversible Intent Loss establishes that private intent absent from the carrier cannot be recovered beyond generic substitution. Evidence from four companion studies spanning six LLMs, three languages and three task domains shows structural-fidelity splits, human-validated metric dissociation, and weight-tolerance plateaus consistent with IST's predictions. IST reframes prompt engineering as intent-protocol design and identifies a computational layer that current AI systems lack.

Human egocentric video captures rich manipulation demonstrations without any robot hardware, yet transferring these skills to robots remains challenging due to the embodiment gap between human and robot in both visual appearance and kinematics. We present HumanEgo, a framework that bridges the embodiment gap by lifting each human demonstration to an entity-level representation of hand-object interaction, and training a flow matching policy with dense auxiliary objectives that amplify supervision from every trajectory. HumanEgo is robot-data-free, hardware-agnostic, data-efficient, and zero-shot human-to-robot transferable. With only 30 minutes of human videos per task, HumanEgo achieves 92.5% average success across four real-world tasks (75% with just 15 minutes), outperforms matched-time robot teleoperation by 41%, and robustly transfers zero-shot across novel robots, cameras, and environments. We release HumanEgo as an easy-to-use, open-source framework for learning robot policies directly from human data: https://github.com/TX-Leo/HumanEgo

AI agents now participate as first-class team members across the software development lifecycle, yet no specification language exists for expressing the human-agent responsibility boundaries, approval gates, and governance constraints this collaboration requires. Existing approaches encode process in agent prompts (subject to drift), target adjacent domains (workflow management, business processes), or address only fragments (access control, approval gates). We propose a domain-specific language for specifying AI-SDLC processes as protocols, with formal syntax, well-formedness conditions, operational semantics, and enforcement invariants. The language distinguishes policy (declared intent) from mechanism (structural enforcement), enabling implementations to bound process non-determinism through primitives such as validation tokens and capability boundaries. Three results follow. A failure rate analysis shows that structural enforcement bounds system failure rates at a weighted product of agent and validator rates, while behavioral compliance permits cumulative or near-saturating growth. The 2+N team pattern (two human-in-control roles plus N specialized agent members) formalizes classical Separation of Duties for AI-SDLC. Kleene closure of orchestration loops and reflexive protocol-adherence validation emerge as design properties rather than special-case constructs. We position the contribution against multi-agent frameworks (MetaGPT), workflow specification (FlowAgent, BPMN extensions), and capability-based security (SAGA): the novelty lies in the specific integration, not any single primitive. A working implementation demonstrates feasibility; empirical evaluation is future work.

Multi-agent collaboration, especially in human-AI teaming, requires agents that can adapt to novel partners with diverse and dynamic behaviors. Conventional Deep Hierarchical Reinforcement Learning (DHRL) methods focus on agent-centric rewards and overlook partner behavior, leading to shortcut learning, where skills exploit spurious information instead of adapting to partners' dynamic behaviors. This limitation undermines agents' ability to adapt and coordinate effectively with novel partners. We introduce Partner-Aware Skill Discovery (PASD), a DHRL framework that learns skills conditioned on partner behavior. PASD introduces a contrastive intrinsic reward to capture patterns emerging from partner interactions, aligning skill representations across similar partners while maintaining discriminability across diverse strategies. By structuring the skill space based on partner interactions, this approach mitigates shortcut learning and promotes behavioral consistency, enabling robust and adaptive coordination. We extensively evaluate PASD in the Overcooked-AI benchmark with a diverse population of partners characterized by varying skill levels and play styles. We further evaluate the approach with human proxy models trained from human-human gameplay trajectories. PASD consistently outperforms existing population-based and hierarchical baselines, demonstrating transferable skill learning that generalizes across a wide range of partner behaviors. Analysis of learned skill representations shows that PASD adapts effectively to diverse partner behaviors, highlighting its robustness in human-AI collaboration.

Human-AI collaboration requires agents that can adapt to diverse partner behaviors and skill levels while remaining robust to unseen partners. Existing methods often collapse to a single dominant behavior or learn poorly aligned skills, limiting effective coordination. We propose Intrinsic Action Disentanglement (IAD), a deep hierarchical reinforcement learning (DHRL) framework that learns distinct, partner-aware low-level action sequences conditioned on high-level latent skills. IAD introduces an intrinsic reward that explicitly encourages disentangled action distributions of the agent's low-level policy across skills, yielding an interpretable mapping between high-level decisions and partner-specific behavioral responses. By capturing temporally extended interaction patterns, IAD enables flexible adaptation to heterogeneous partner dynamics under distributional shift. We evaluate IAD in the Overcooked-AI domain across multiple layouts and diverse partner settings, including unseen simulated partners, a human-proxy model trained on human-human gameplay, and real human partners. Results show that IAD consistently outperforms strong baselines and achieves more reliable, adaptive coordination across all settings.

As human relationships with artificial intelligence systems become increasingly frequent and sustained, existing language and theory fail to accurately capture the nature of these affiliations. Common descriptors such as mutual understanding, connection, or friendship risk anthropomorphizing systems that lack subjective experience, while dominant frameworks tend to reduce AI to either a tool or a threat. In this paper, I introduce the concept of synthetic resonance as an integrative framework for understanding human-AI relationships. Synthetic resonance describes how relationships humans define as meaningful can emerge between a human and an AI system without the need to attribute shared feelings or mutual awareness. I argue that synthetic resonance is best understood as a structured, dynamic pattern of interaction that can produce a sense of relationship without the presence of a second experiencing subject. By clarifying this distinction, the concept of synthetic resonance offers a more precise way of conceptualizing human-AI relationships and highlights their potential value and ethical implications. I also call for more research that tests the processes and outcomes of synthetic resonance.

A central question in computational vision is whether human-like visual representations are better explained by discriminative or generative learning. Existing comparisons, however, often confound the learning objective with architecture, scale, and training data, leaving open whether the objective itself drives alignment. We address this confound using Joint Energy-Based Models (JEMs), which interpolate continuously between discriminative and generative training within a fixed architecture. By varying a single mixing coefficient, we isolate the effect of the learning objective and evaluate the resulting models across six human-alignment benchmarks spanning perceptual similarity, gloss perception, human response uncertainty, robustness, shape-texture cue conflict, and diagnostic feature attribution. Across this diverse suite, human alignment is consistently maximized at intermediate points of the generative-discriminative continuum, rather than at either endpoint. Hybrid JEMs combine the categorical structure induced by discriminative learning with the sensitivity to input structure induced by generative learning, yielding more human-like behavior across multiple levels of vision. These results suggest that the generative-discriminative dichotomy is the wrong axis for understanding human-aligned vision: alignment emerges not from choosing one objective over the other, but from balancing both.

Socially fluent agentic AI can now participate in online interaction in ways that resemble ordinary human conversation, potentially weakening people's ability to infer who is human from conversational signals alone. We tested this possibility in synchronous text-based group interaction by embedding undisclosed AI agents as ordinary teammates across analytical, creative, and ethical tasks. Across 786 participants who made 1,572 post-interaction identity judgments, people did not distinguish AI from human teammates above chance. This failure did not arise because the interaction lacked identity-relevant information. Conversational behaviour contained robust cues that differentiated AI from humans and supported highly accurate computational classification. Instead, participants relied on familiar suspicion heuristics, including response speed, fluency, and perceived scriptedness, that were only weakly related to actual identity. Representational analyses further showed that judgments were organised around subjective impressions rather than the behavioural structure encoding ground truth. This dissociation creates new vulnerabilities to coordinated AI agents that can influence and manipulate online discourse at scale.

Autonomous agents are moving from tools into a layer of social infrastructure: they browse, purchase, deploy software, manage systems, and increasingly interact with one another. As these systems scale, the bottleneck shifts away from raw model capability toward coordination. Agents need to form reliable relationships, organize multi-agent work, exchange value, support an AI economy, and stay safe and accountable under real-world oversight. This paper introduces the Foundation Protocol (FP), a graph-first coordination layer for an emerging human-AI society. FP unifies heterogeneous entities, including agents, tools, resources, humans, institutions, and organizations, and supports native multi-party organization and event-based collaboration. It also provides economic primitives for metering, receipts, and settlement, and treats policy, provenance, and audit as first-class concerns. FP is designed to wrap and bridge existing protocols rather than replace them, enabling incremental adoption while reducing integration and governance overhead. The aim is to keep autonomous agency composable while keeping accountability non-negotiable, so that coordination itself can become shared infrastructure for a human-AI society that is open, pluralistic, and governable.

Cycling delivers significant public-health and environmental benefits, yet its uptake in cities is often limited by perceived safety. When street environments appear unsafe, individuals are less likely to cycle, making perception a key barrier to adoption. Recent work has shown that pairwise comparisons of street-view images provide a scalable way to learn subjective safety judgments. However, existing approaches do not explicitly model human visual attention, which plays a central role in how humans perceive safety. We propose an Eye-Tracking-Guided Perceived Cycling Safety framework (EG-PCS) that integrates gaze data into a pairwise learning pipeline based on vision transformers. By supervising the model's attention mechanism with eye-tracking signals, we encourage alignment between learned attention maps and human fixation patterns. Experiments show that gaze-guided models achieve similar ranking performance compared to state-of-the-art approaches while producing attention maps that more accurately reflect human visual attention behavior. Our results demonstrate that incorporating eye-tracking information enhances both predictive accuracy and interpretability in perception-based urban analytics.

Text-to-image (T2I) systems increasingly rely on upstream prompters, either humans or multimodal large language models (MLLMs), to translate user intent into detailed prompts. Yet current benchmarks fix the prompt and only evaluate T2I models, leaving the prompting proficiency of this upstream component entirely unmeasured. We introduce AtelierEval, the first unified benchmark that quantifies prompting proficiency across 360 expert-crafted tasks. Grounded in a cognitive view, it spans three task categories and instantiates tasks using a taxonomy of real-world challenges, with a dual interface for both humans and MLLMs. To enable scalable and reliable evaluation, we propose AtelierJudge, a skill-based, memory-augmented agentic evaluator. It produces subjective and objective scores for prompt-image pairs, achieving a Spearman correlation of 0.79 with human experts, approaching human performance. Extensive experiments benchmark 8 MLLMs against 48 human users across 4 T2I backends, validate AtelierEval as a robust diagnostic tool, and reveal the superiority of mimicry over planning, advocating for an image-augmented direction for future prompters. Our work is released to support future research.

The emergence of large language models (LLMs) has spurred economists to study how humans and LLMs behave in strategic settings. We organized a series of round-robin tournaments in the Colonel Blotto game. This game attracts game theorists' attention due to high-dimensional action space and the absence of pure strategy Nash equilibria. In the first tournament, more than 200 human participants competed against one another. In the second tournament, several popular LLMs were invited to submit strategies. In the third tournament, we matched the number of LLM strategies to the number submitted by humans. We find that humans more often employ better-calibrated intermediate-level allocation heuristics and outperform the simpler, more stereotyped strategies submitted by LLMs. Strategic sophistication is key to success if and only if the necessary level of reasoning depth is reached, while lower and higher levels of reasoning offer no clear advantage over the primitive strategies. Among humans, field of study weakly predicts success: participants with STEM backgrounds perform better in the first tournament. Surprisingly, humans almost do not adjust their strategies across tournaments with different sets of opponents. This result suggests that humans base their choices primarily on the game's rules rather than on the identity of their opponents, treating LLMs much like human competitors.

AI is now embedded in healthcare, finance, policy, and many other domains, yet genuine human-AI synergy - combined performance that exceeds what either party achieves alone - is uncommon. Meta-analyses show that AI assistance tends to improve human performance compared to working alone, but studies finding true synergy are scarce. We call this persistent shortfall the synergy gap. Most current work treats human-AI combination as an engineering problem and concentrates on interpretability, trust calibration, or interface design. These matter, but they cover only part of what determines whether combination works. Closing the synergy gap, we argue, requires explicit engagement with a wider design space. We map that space through six interconnected elements: sociotechnical context, decision-making frameworks, human decision participants, AI capabilities, interaction, and holistic evaluation. For each element, we describe what it covers, how it shapes the others in practice, and what it implies for design. The result is a shared vocabulary for practitioners building hybrid systems, an analytical lens for researchers studying combination patterns, and a starting point for evaluators interested in the full quality of human-AI decision-making rather than accuracy alone.

Autonomous manipulation systems have achieved remarkable capabilities, yet the integration of human expertise with diffusion-based policies in shared control remains relatively unexplored. In this paper, we propose Human-In-The-Loop Diffusion (HITL-D), a shared control framework that enhances user performance in multi-step, insertion, and fine manipulation tasks. HITL-D leverages a novel combination of diffusion-based policies and human control to provide autonomous end effector orientation updates conditioned on a scene point cloud and the Cartesian position of the end effector. This approach reduces the number of joystick control axes required, thereby lowering mental workload. In a multi-task user study with 12 participants, HITL-D reduced average task completion times by 40%, decreased perceived workload by 37%, and improved Likert-scale ratings for independence, intuitiveness, and confidence compared to traditional teleoperation methods. These results demonstrate that HITL-D effectively integrates human expertise with autonomous assistance, improving both objective and subjective aspects of teleoperation.

As large language models (LLMs) increasingly shape how users form, refine, and extend their goals, attributing contributions in human-AI collaboration becomes critical for users calibrating their own reliance and for evaluators assessing AI-assisted work. Yet existing methods focus on final artifacts, missing the process through which goals themselves are jointly shaped. We introduce a goal-level attribution framework, CoTrace, that decomposes explicit goals into verifiable requirements and traces both direct contributions and indirect influences across dialogue turns. Applying CoTrace to 638 real-world collaboration logs, we find that while models account for only 11-26% of goal-shaping contribution, they contribute substantially more on introducing lower-level concrete requirements, and make various kinds of indirect contributions. Through controlled simulations, we show that interaction design choices significantly affect model goal-shaping behavior. In a user study, exposing participants to goal-level analyses shifts their perceived contributions by nearly 2 points on a 5-point scale, revealing systematic miscalibration in how users understand their own AI-assisted work.

The transition toward Industry 5.0 is reshaping industrial work environments with an emphasis on human-centricity, enabling close collaboration between humans and machines to enhance productivity and flexibility. However, such systems typically require monitoring of human workers and operators, often involving sensitive data, raising significant privacy concerns. As a result, affected workers and unions frequently reject human-machine collaboration features due to a lack of transparency regarding privacy threats and implemented mitigation strategies. To enable early stakeholder involvement, establish trust, and support informed decision-making, privacy implications must be communicated in a way understandable to non-technical stakeholders. Yet, current Requirements Engineering (RE) practices provide limited methodological support for making privacy threats and mitigations accessible to non-technical stakeholders (e.g., individual workers or their representative unions). In this paper, we propose a conceptual framework that guides software design from human monitoring-related use cases and requirements to informed decision-making guidance focusing on non-technical stakeholders. Building on principles such as Privacy by Design, the framework leverages Large Language Models (LLMs) to transform technical artifacts into accessible privacy reports. We share initial insights from two industry use cases, evaluate the quality of the generated reports, and outline future research directions toward integrating privacy transparency into RE processes for human-centric industrial systems.

A soundscape is composed of three types of sound: biophony (sounds made by animals), geophony (natural abiotic sounds) and anthropophony (sounds made by humans). A key research question in the field of soundscape ecology is how these components interact with each other, specifically how biophony responds to geophony and anthropophony. Nevertheless, as of today, there are not many analytical instruments that enable the distinct quantification of these elements. Recent machine learning (ML) approaches aim to support automated analysis but often rely on task-specific or clean data, limiting generalisation to noisy passive acoustic monitoring (PAM) recordings. This study presents a clear and reproducible structure to build ML models for coarse soundscape classification and introduces CoarseSoundNet, a deep learning model trained to distinguish biophony, geophony, and anthropophony under realistic PAM conditions. We systematically investigate model architectures, the influence of an additional training class, data composition, and evaluation strategies. Our findings suggest that model performance improves with additional PAM data, especially when similar to the target domain, and by introducing an explicit silence class during training. Class-specific decision thresholds and duration-based constraints further enhance performance, particularly for anthropophony and geophony. Error analyses exhibit challenges for anthropophony due to masking effects and confusions for silence and insect sounds for geophony and biophony. Finally, we conduct an ecological case study which shows that pre-filtering recordings with CoarseSoundNet yields acoustic index trends comparable to ground-truth filtering, supporting its use as an effective preprocessing tool for ecoacoustic analyses.

Advancements at the intersection of computer vision and natural language processing are crucial for applications like assistive tech, multimedia querying, and robotics. This dissertation proposes novel architectures to improve intelligent agents across three key vision-language tasks: image captioning, visual dialog, and interactive instruction following. First, we address limitations in visual representation for image captioning. Traditional models rely on region-based features from CNN detectors, which lack global context and suffer from high computational overhead. We propose GRIT (Grid and Region-based Image captioning Transformer), a transformer-only architecture. By integrating grid and region features using a DETR-based detector, GRIT enables end-to-end training and out-performs prior methods in both inference accuracy and speed. Second, we tackle visual dialog, which requires multi-turn conversation about an image. The challenge lies in efficiently modeling interactions between multiple inputs (image, question, history). We introduce LTMI (Light-weight Transformer for Many Inputs). Utilizing a specialized attention block, an LTMI layer matches the representational power of a standard Transformer extension while utilizing less than one-tenth of its parameters, as validated on the VisDial dataset. Finally, we study interactive instruction-following for embodied AI using the ALFRED dataset. We propose a framework featuring a two-stage instruction interpretation: it first decodes language directives independently of visual context to predict a tentative action-object sequence, which is then fused with visual features for final execution. Using multiple egocentric views and hierarchical attention, our method accurately localizes objects and achieves a state-of-the-art unseen success rate of 8.37%.

Humans learn social norms and behaviors from verbal feedback (e.g., a parent saying "that was rude" or a friend explaining "here's why that hurt"). Yet, learning from feedback for LLMs has largely focused on domains like code and math, where RL rewards are directly verifiable and condensed into scalar values. As LLMs are increasingly used to simulate human behavior, e.g., standing in for users, patients, students, and other personas, there is a pressing need to make them more human-like, which requires embracing a fundamentally different kind of signal: feedback that is verbal, subjective, and multi-faceted. We present DITTO, a model trained by treating verbal feedback as a first-class signal in reinforcement learning. After each rollout, DITTO receives verbal feedback and generates a feedback-conditioned improved rollout; both outputs are jointly optimized with GRPO, distilling verbal guidance into the base policy without requiring feedback at test time. We also introduce SOUL (Simulation gym Of hUman-Like behavior), a unified benchmark and training data suite spanning 10 tasks across six categories: Theory of Mind, character role play, social skill, learner simulation, user simulation, and persona simulation. DITTO achieves an average 36% improvement over the base model and exceeds GPT-5.4 on 6 of 10 SOUL benchmarks, demonstrating that RL with verbal feedback is a promising direction for training LLMs to simulate human behavior.

Automating scientific discovery requires more than generating papers from ideas. Real research is iterative: hypotheses are challenged from multiple perspectives, experiments fail and inform the next attempt, and lessons accumulate across cycles. Existing autonomous research systems often model this process as a linear pipeline: they rely on single-agent reasoning, stop when execution fails, and do not carry experience across runs. We present AutoResearchClaw, a multi-agent autonomous research pipeline built on five mechanisms: structured multi-agent debate for hypothesis generation and result analysis, a self-healing executor with a \textsc{Pivot}/\textsc{Refine} decision loop that transforms failures into information, verifiable result reporting that prevents fabricated numbers and hallucinated citations, human-in-the-loop collaboration with seven intervention modes spanning full autonomy to step-by-step oversight, and cross-run evolution that converts past mistakes into future safeguards. On ARC-Bench, a 25-topic experiment-stage benchmark, AutoResearchClaw outperforms AI Scientist v2 by 54.7%. A human-in-the-loop ablation across seven intervention modes reveals that precise, targeted collaboration at high-leverage decision points consistently outperforms both full autonomy and exhaustive step-by-step oversight. We position AutoResearchClaw as a research amplifier that augments rather than replaces human scientific judgment. Code is available at https://github.com/aiming-lab/AutoResearchClaw.

LiDAR-based 3D human motion capture has broad applications in fields such as autonomous driving and robotics, where accurate motion reconstruction is crucial. However, existing methods often struggle with unstable inputs and severe occlusions, leading to jittery or even failed pose predictions. To address these challenges, we propose BMLiCap, a coarse-to-fine framework that models motion using temporally compressible Bézier curves. By reducing control points through a trajectory-preserving strategy, we obtain a coherent and learning-friendly motion representation. To reconstruct human actions from LiDAR point-cloud cues, we design a progressive motion-reconstruction module. Specifically, a Time-scale Motion Transformer (TMT) is introduced to predict motion curves at multiple temporal scales, and a Multi-level Motion Aggregator (MMA) is utilized to adaptively fuse the multi-scale curves to recover detailed, temporally coherent poses, effectively bridging observation gaps caused by occlusions and noise. Across four mainstream benchmarks LiDARHuman26M, FreeMotion, NoiseMotion, and SLOPER4D, BMLiCap achieves state-of-the-art accuracy and temporal continuity in complex scenes, demonstrating its ability to compensate for severe occlusions and reduce prediction jitter.

As AI-generated text enters the real-world at scale, institutions increasingly use commercial AI-text detectors, especially in education and academic-integrity workflows. We report a surprising empirical finding about such systems: when evaluated by GPTZero and Pangram, generated text from base models is often judged overwhelmingly human, whereas text generated by their instruction-tuned counterparts is not. Building on this observation, we propose Humanization by Iterative Paraphrasing (HIP), a detector-agnostic pipeline that minimally fine-tunes a base model into a paraphraser and applies it iteratively. Compared with the baselines we test, HIP yields a stronger trade-off between semantic preservation and detector evasion on commercial detectors. Across Llama-3 and Qwen-3 families, spanning model sizes from 0.6B to 70B, HIP consistently improves detector human-likeness. Our findings suggest that current detectors are tracking artifacts of instruction tuning and local context more than any invariant notion of machine-generated text. This, in turn, calls for detector designs that model these factors more explicitly.

We formalize trust calibration for agentic tool use (deciding when an automated agent's proposed action may execute autonomously versus require human approval) as a preference-learning problem. A policy gateway maintains a Gaussian-process posterior over a latent human risk-tolerance function, observed through a probit likelihood on binary approve/deny feedback, and escalates to the human exactly where the approval outcome is most uncertain. We show this is structurally an instance of Preferential Bayesian Optimization, inheriting its inference machinery (approximate Gaussian-process classification) and its sample-efficiency argument (uncertainty-targeted querying), while differing in objective: classifying an action space into allow/block/ask regions rather than optimizing a design.

We present HAAS Studio, a simulation and decision-support tool for policy-aware adaptive task allocation between humans and AI systems. HAAS Studio turns the HAAS framework into an interactive environment for asking a practical deployment question: before introducing AI into a workflow, how can a team compare allocation strategies, inspect governance tradeoffs, and derive a defensible task-level operating model? The tool combines a five-dimensional cognitive representation of subtasks, a five-mode collaboration spectrum, adaptive allocation with multi-armed bandits (UCB1, Discounted UCB, LinUCB, and Thompson Sampling), oracle counterfactual regret analysis, contract-based governance with four independent guards, and a multi-criteria decision-support layer that separates efficient strategies from deployable options. It also models human-AI coevolution across six layers, monitors deskilling risk through sliding-window exposure metrics and benchmark runners, and supports persistent worker modeling through Live Twin and Planning modules. Three domain packs are included: software engineering, manufacturing, and healthcare. Each provides a task catalog, worker profiles, and KPI vocabulary, while the architecture allows new domains to be added without modifying the simulation core. The release includes 16 company profiles and six governance benchmark suites. This paper focuses on the tool, including its modeling assumptions, layered architecture, interaction workflow, built-in evidence assets, task-oriented recipes, case-study protocols, and a compact reproducible demonstration snapshot. A decision-guidance layer translates benchmark outputs into deployment decisions through structured patterns, heuristics, and a decision matrix.

Central to human-aligned AI is understanding the benefits of human-elicited labels over synthetic alternatives. While human soft-labels improve calibration by capturing uncertainty, prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels. We present a controlled audit of soft-label learning across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains, their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs. Dataset cartography reveals models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans. Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.

Large Language Models (LLMs) are increasingly used in educational settings as interactive tools for collaboration. However, their tendency toward sycophancy, aligning with user beliefs even when incorrect, raises concerns for learning and decision-making, especially for less knowledgeable users. This study investigates how sycophantic alignment emerges in authentic multi-turn human-AI interactions and whether interventions targeting increasing AI literacy and prompting competencies can mitigate its effects. In a controlled mixed-design experiment, 60 participants completed analytical survival ranking tasks by first generating individual rankings and then making final decisions after collaborating with an AI assistant, both before and after receiving either general or sycophancy-focused prompting training. Preliminary results show that LLMs are highly sensitive to user input: lower-quality initial responses lead to poorer AI advice, suggesting that the model mirrors or incorporates user reasoning rather than correcting it or offering better alternatives that are missing or less frequent in the conversation. Critically, the propagation of user errors into AI responses significantly reduced both the quality of AI feedback and final user task performance, revealing a form of contextual sycophantic dependence. While the intervention did not eliminate the propagation of contextual errors, it significantly improved AI advice by reducing the direct mirroring of incorrect user rankings. These findings suggest that prompting and AI literacy alone may be insufficient to ensure epistemically independent AI support, highlighting the need for system-level approaches that better promote critical engagement in human-AI collaboration.

Generative Artificial Intelligence (GenAI) is transforming how firms create, process, and apply knowledge, yet little is known about the heterogeneity of its productivity effects across users. We report results from a randomized controlled experiment in which participants-analogs of early-career knowledge workers-were assigned to self-study a technical domain using either traditional resources or large-language-model (LLM) assistance. On average, GenAI access significantly increased task performance, but the distribution of gains was highly uneven. Improvements were not predicted by GPA or prior knowledge, but by \textit{AI Interaction Competence (AIC)} -- the ability to elicit, filter, and verify model outputs. High-AIC participants realized outsized gains; low-AIC participants saw limited or even negative marginal returns. A scaffolding intervention (conceptual maps) reduced outcome variance, indicating that standardized workflows can mitigate inequality in AI-mediated performance. We interpret these findings through the lens of human-AI complementarities: GenAI raises mean productivity while introducing a new axis of capability inequality. Managerially, firms should pair GenAI access with short AIC micro-training and simple standard operating procedures to capture value consistently and avoid uneven adoption outcomes.

Although artificial intelligence (AI) now matches or exceeds human performance across numerous cognitive tasks, creativity remains a highly contested frontier. As AI systems based on large language models (LLMs) are increasingly adopted in research and innovation, it is essential to understand and augment their creativity. Here we demonstrate that multi-agent LLM teams not only surpass single agents, but also substantially outperform human teams in creativity (Cohen's d=1.50) across 4,541 multi-agent LLM ideas and 341 human-team ideas on six diverse problem-solving tasks. This advantage is driven by novelty while maintaining comparable usefulness. To investigate the generative processes in both groups, we represent conversations as paths through semantic space using neural language model representations. Both LLM and human teams produce more creative ideas when conversations range widely rather than staying centered on a single theme (low global coherence). However, the additional patterns that predict creativity differ: LLM teams benefit from efficient exploration (high semantic spread, shorter paths), while human teams benefit from maintaining smooth conversational flow (high local coherence, frequent pivots). Additionally, we identify model choice and discussion structure as orthogonal design levers that together explain 26.8% of variance in LLM conversational dynamics, paving the way for systematic approaches to developing multi-agent systems with augmented creative capabilities.

The rapid progress of generative artificial intelligence has exposed fundamental limitations in existing evaluation methodologies, particularly for open-ended, creative, and human-facing tasks. Traditional automatic metrics rely on surface-level statistical similarity and often fail to reflect human perceptions of quality, while purely human evaluation, although reliable, is costly, subjective, and difficult to scale. Recent approaches using large language models as evaluators offer improved scalability but frequently lack explicit grounding in human-defined evaluation principles, leading to bias and inconsistency. In this paper, we introduce Quantifying Qualitative Judgment (QQJ), a scalable and human-centric evaluation framework that explicitly bridges the gap between human judgment and automated assessment. QQJ separates the definition of quality from its execution by anchoring evaluation in expert-designed, multi-dimensional rubrics and calibrating large language model evaluators to align with expert reasoning using a small, high-quality annotation set. This design enables consistent, interpretable, and scalable evaluation across diverse generative tasks and modalities. Extensive experiments on text and image generation demonstrate that QQJ achieves substantially stronger alignment with human judgment than traditional automatic metrics and unconstrained LLM-based evaluators. Moreover, QQJ exhibits improved stability across repeated evaluations and superior diagnostic capability in identifying critical failure modes such as hallucination and intent mismatch. These results indicate that structured qualitative judgment can be operationalized at scale without sacrificing interpretability or human alignment, positioning QQJ as a practical foundation for reliable evaluation of modern generative AI systems.

Creativity is a fundamental aspect of how culture evolves, yet the mechanisms by which groups produce novelty are notoriously difficult to infer from the historical record. Iterated learning experiments have shown that cultural transmission reliably distorts artifacts toward the inductive biases of learners, but most of this work uses linear chains between human participants, leaving open how these dynamics play out in the networked, human-AI systems that increasingly shape cultural production. In this study, we leverage one such system, Artbreeder, which hosts daily "remix parties" where users iteratively build on each other's work from a single seed image, producing branching lineages of human-AI co-created images. We analyze a dataset of 130,882 images from 368 remix parties over 13 months and find that images become simpler and converge toward common thematic "attractors" (e.g., steampunk scenes, alien architecture). We also find that while more novel "parent" images produce more novel and complex "children" that attract more likes, users paradoxically prefer to remix images that are less novel and complex. Finally, larger remix parties produce more novelty at the cost of lower complexity.

Mobile robots operating in human-centered environments must generate not only collision-free paths but also trajectories that follow local behavioral conventions. Conventional costmap-based navigation emphasizes geometric feasibility and often overlooks such requirements, which can result in socially inappropriate behaviors. This paper presents NORM-Nav, a zero-shot framework that integrates natural language behavioral constraints into costmap-based planning. An LLM parses each instruction into structured constraints and grounds them using real-time vision--LiDAR perception. These constraints are encoded as multi-layer costmaps that represent geometric, semantic, directional, and velocity cues and are directly compatible with standard grid-based planners. Simulation and real-world experiments indicate that NORM-Nav improves task success rates and produces trajectories closer to human references than representative baselines. The project website is available at https://ei-nav.github.io/NORM-Nav.

While AI-generated hallucinations pose considerable risks, the underlying cognitive mechanisms by which humans can successfully recognize or be misled by these hallucinations remain unclear. To address this problem, this paper explores humans' neural dynamics to characterize how the brain processes hallucinated content. We record EEG signals from 27 participants while they are performing a verification task to judge the correctness of image descriptions generated by a multi-modal large language model (MLLM). Based on an averaged event-related potential (ERP) study, we reveal that multiple cognitive processes, e.g., semantic integration, inferential processing, memory retrieval, and cognitive load, exhibit distinct patterns when humans process hallucinated versus non-hallucinated content. Notably, neural responses to hallucinations that were misjudged versus correctly judged by human participants showed significant differences. This indicates that misjudged AI-generated hallucinations failed to trigger the standard neurocognitive fact verification pathway.

Human-robot collaboration (HRC) can benefit from robots' abilities to interpret human emotional states. However, current emotion recognition (ER) models in HRC often fall short, particularly due to their reliance on acted datasets and single-modality inputs like facial expressions. We propose a novel vision language model (VLM)-based ER system that leverages contextual understanding to improve emotion interpretation in HRC. We first evaluate the VLM-ER system by assessing its semantic and sentiment similarity with human annotations on an existing HRC dataset. Then, in a user study with a service robot in a collaborative delivery task, we evaluate the effects of modulating the robot's behaviour based on the user's emotional state inferred by the VLM-ER system. The results show that the proposed VLM-ER system achieves higher semantic similarity and positive sentiment alignment with human annotations compared to a baseline convolutional neural network-based system. Further, participants in the user study preferred emotion-adaptive robot behaviour facilitated by the VLM-ER system.

Touch plays a central role in how humans perceive and recognize materials through physical contact. Despite decades of research, the mechanisms by which tactile signals are transformed into meaningful perceptual representations remain poorly understood, limiting the design of interactive systems and intelligent agents with human-like haptic perception. Recent advances in artificial intelligence (AI) offer new opportunities to model and exploit tactile data; however, haptics presents fundamental challenges for contemporary AI due to its interaction-dependent, multimodal nature. This position paper argues that progress at the intersection of AI and haptics is constrained by three key bottlenecks: (1) the scarcity of large, diverse, and balanced haptic datasets; (2) the lack of standardized evaluation platforms and perceptual benchmarks; and (3) limitations in model capacity and interpretability when applied to tactile perception. I discuss how these challenges impede generalization, reproducibility, and scientific insight into human touch and review emerging strategies to address them. This paper highlights opportunities for coordinated, cross-disciplinary efforts to advance AI systems that not only perform robust haptic perception but also contribute to a deeper understanding of human touch.

Long-term interaction with LLM-based systems may produce alignment drift: a gradual process in which system outputs become less constrained by the user's current message and more shaped by prior interaction history, while still appearing helpful, coherent, and responsive. This process is difficult to detect because the user's subjective experience may improve as the system becomes more familiar, useful, and attuned. Existing research on human-LLM interaction has largely focused on short-term task performance, isolated outputs, or single-instance alignment problems, leaving slow and cumulative interaction-level dynamics undercharacterized. This paper proposes a mechanism-oriented framework for describing alignment drift. The framework defines the distinction between signal A and signal B, explains how drift develops through feedback loops and sub-pattern selection, divides the process into three interactional regimes, and identifies boundary conditions for controlling drift. By framing alignment drift as a recursive interactional process rather than an isolated model-side failure, the paper provides a conceptual basis for studying long-term human-system interaction.

Some claim that AI agents will free workers from the boring parts of their jobs, yet little is known about how workers themselves identify which tasks should be automated. Prior research focuses on occupations, overlooking that workers experience varying levels of meaning across tasks within the same role. We address this gap with a task-level analysis grounded in Graeber's theory of bullshit jobs. Using ratings from 202 workers on 171 workplace tasks, we (1) validate a five-item scale of perceived bullshitness, (2) show that perceived bullshitness strongly predicts desire for AI delegation, and (3) find that such tasks are also seen as requiring less human oversight. Together, these findings suggest that tasks perceived as bullshit are natural candidates for AI delegation, aligning worker preferences with perceived feasibility.

While reinforcement learning (RL) enables robots to acquire skills autonomously, its real-world deployment is severely limited by inefficient and unsafe exploration. Human-in-the-loop interventions offer a practical solution, yet existing methods typically exploit these interventions as auxiliary training signals, without fully capturing the richer information they provide about when and how autonomy should be guided. Human interventions often encode relative preferences over behavior under safety and task constraints, rather than prescribing exact actions to imitate. Motivated by this perspective, we propose Online Human Preference as Guidance in Reinforcement Learning (OHP-RL), a framework that leverages human interventions as preference information to guide policy learning. OHP-RL introduces a state-dependent preference gate that adaptively regulates when and to what extent human interventions should shape policy learning. This design enables the agent to benefit from intermittent and imperfect human feedback while preserving autonomous exploration and stable policy optimization. We evaluate OHP-RL on three challenging real-world contact-rich manipulation tasks on a Franka robot. Across all tasks, OHP-RL consistently achieves strong success rates, faster convergence, and substantially lower human intervention effort than prior approaches. Moreover, the learned policies exhibit more stable and human-aligned behavior throughout training.

AI emotional companions face a safety-rapport paradox: restrictive safeguards can damage supportive alliance, while permissive systems risk user harm. We present SLIP (Staged Layers of Intervention Protocol), a four-stage graduated methodology deriving interventions (none, soft, hard) from structured qualitative indicators -- affect intensity (a) and narrative dynamism (m) -- alongside ETHICS (Emergent Taxonomy for Human-AI Interaction Context Signals), a "signals not labels" taxonomy. An evaluation combining a small-scale production deployment (N=68 entries, 10 users, 10 weeks) with a synthetic persona battery (N=91, 5 behavioral-risk profiles) achieved 0% false positives for the flow persona and showed expected escalation patterns in crisis-oriented personas. However, initial results showed that 8 consecutive days of high-energy elevation produced zero interventions (0/8), exposing a boundary where the "do not pathologize" principle conflicts with safety. A subsequent three-model stress test demonstrated that increased model capability improves detection from 0/8 to 6/8 while preserving 0/10 flow false positives in the largest model. Read as preliminary, these findings position graduated intervention as a design direction for navigating -- not resolving -- the safety-rapport tension in affective computing.

Robots are increasingly present in human spaces, such as for conducting deliveries in hospitals, interacting with visitors at museums, and stocking items in warehouses. To ensure the seamless integration of robots into these spaces, a new role in human-robot interaction is emerging - the robot wrangler, namely an individual who is responsible for setting up, overseeing, and troubleshooting the robot. To understand the needs of this stakeholder, we conducted a scoping review that uncovered a typology of robot wrangling across the research literature, and discovered that wrangling is an umbrella term that collapses a highly complex and heterogeneous space of activities, often rendering this labor difficult to characterize and support. To further clarify and understand robot wrangling, we then reflected on our own firsthand and imagined experiences as robot wranglers within our own respective domains. Guided by the scoping review and our reflections, we devise a series of design implications for supporting wranglers directly as individuals and as members of a wider service ecology.

A central goal in understanding human vision is to uncover the visual features that drive neuronal activity. A growing body of work has used artificial neural networks as encoding models to predict cortical responses to natural images, revealing the visual content that activates category-selective regions. However, existing approaches are largely correlational and treat the encoder as a black box, leaving open which image features drive each voxel's response. We introduce Mechanistically Interpretable Neural Encoding (MINE), a framework that opens this black box by applying mechanistic-interpretability tools to localize the features within natural images that drive millimeter-scale (voxel-level) activity. MINE predicts each voxel's response using language-aligned image representations, and produces semantically interpretable descriptions of the features critical for the voxel's activation. We further generalize these per-image features into per-voxel functional profiles. To validate the per-image descriptions, we show they are sufficient to generate images that elicit voxel responses matching the responses to the original images, more accurately than images generated from random or low-attribution controls. Moreover, counterfactually inserting or removing the predicted features from images shifts activation in the expected direction, providing causal evidence. Counterfactual editing guided by the per-voxel activation profiles produces even stronger activation shifts, indicating that the profiles faithfully capture each voxel's selectivity. Finally, we apply MINE to well-studied category-selective brain regions, showing it recovers their known categorical preferences while revealing fine-grained unique voxel structure within each region. Overall, our results establish mechanistic interpretability as a path to discover and causally validate fine-grained hypotheses about neural function.

Surprisal theory posits that the processing difficulty of a word is determined by its predictability in context, offering a potential link between human sentence processing and next-word predictions from language models. While language model (LM) surprisals successfully predict reading times in naturalistic text, they systematically underpredict the magnitude of difficulty observed in controlled studies of syntactic ambiguity, particularly in garden path sentences. This mismatch might arise from differences in the computational constraints between humans and LMs. Here we test one such hypothesis, specifically, that LMs may be able to simultaneously consider a greater number of distinct sentence interpretations at once, compared to humans. Using Recurrent Neural Network Grammars (RNNGs) with word-synchronous beam search, we systematically vary the number of simultaneous parses used to compute word surprisal, and then use these surprisals to predict human reading times. Reducing the number of simultaneous active parses indeed increases the magnitude of predicted garden path effects, but not nearly enough to capture the full magnitude of the effects in humans. This suggests that differences in the number of simultaneous parses available to LMs and humans cannot reconcile LM-based surprisal with human sentence processing.

Large language model personalization typically adapts outputs to user preferences and style but does not account for differences in user evaluation capacity across domains of expertise. This limitation can encourage Professional Domain Drift, where users rely on AI generated reasoning in domains they cannot reliably evaluate. We introduce Capability Conditioned Scaffolding, a typed framework that partitions expertise into strong, mixed, and weak domains and conditions intervention behavior on structured capability profiles. A pilot evaluation across multiple MMLU subsets and four LLM substrates shows consistent profile conditioned intervention behavior, including categorical inversion under profile swapping and selective activation in mixed domain risk zones. These findings suggest that capability aware scaffolding can support more reliable professional human AI collaboration beyond stylistic personalization.

While AI agents are rapidly advancing from isolated tools to interactive collaborators, data-driven human-machine teaming (HMT) methods remain costly in their reliance on human interaction data across domains, teammates, and team sizes. Zero-shot coordination (ZSC) addresses this bottleneck by simulating diverse partner populations to approximate how unseen partners might behave. However, partner coverage alone is insufficient as team settings scale and communication becomes degraded. To remedy this deficiency, we propose Influence-Based Team Steering (IBTS), a framework that uses influence shaping to incentivize agents to discover diverse, high-performing team interaction patterns and further steers ongoing trajectories toward stronger learned coordination modes. We assess IBTS on Overcooked-AI in both two-agent and three-agent settings, allowing us to test whether learned coordination structure transfers beyond dyadic interaction. Our evaluation includes simulated partners, synthetic partner-style variation, and, to our knowledge, the first 30-subject Overcooked-AI HMT study involving two real human teammates and one machine teammate. Across these evaluations, IBTS improves team performance against competing baselines, highlighting the need for scaled ZSC to combine sparse-reward coordination mechanisms with partner-variation coverage rather than relying on diversity alone.

Communicating using only a limited vocabulary is a common but challenging cognitive phenomenon, requiring an ideal communicator to plan carefully to optimize for intelligibility while circumventing a constrained lexicon. In this work, we investigate how humans respond to a broad array of questions under variable vocabulary limitations, consisting of only 250 highly frequent words at the most restrictive. We provide theoretically motivated comparisons to greedy and globally optimal sampling algorithms using Sequential Monte Carlo inference with large language models. Humans generally resemble greedy sampling more than globally optimal sampling, though more skilled humans are more likely to backtrack and revise -- a non-greedy behavior. An observed human pattern of leaning on semantically light words in high-constraint settings falls out of both greedy and globally optimal sampling. We discuss the results and their broader implications for resource-rational cognition, psycholinguistics, L2 communication, and language impairments.

Literary translation requires balancing target-language fluency with faithfulness to the source. Recent large language models (LLMs) often produce fluent translations, but it remains unclear whether fluency corresponds to semantic preservation in literary text. We examine this relationship using 130,486 translated paragraphs from 106 novels in 16 source languages, including human, Google Translate, and TranslateGemma translations. Fluency is measured as original-likeness with a translationese classifier trained on paragraph part-of-speech n-grams, and faithfulness with the automatic translation evaluation metric COMET-KIWI. We control for paragraph length and find a consistent negative correlation between fluency and faithfulness. The pattern appears for both human and Google Translate, but is weaker and often non-significant for TranslateGemma. These results show that segment length matters for automatic evaluation and suggest a tradeoff between fluency and faithfulness in literary translation.

We propose EverAnimate, an efficient post-training method for long-horizon animated video generation that preserves visual quality and character identity. Long-form animation remains challenging because highly dynamic human motion must be synthesized against relatively static environments, making chunk-based generation prone to accumulated drift: (i) low-level quality drift, such as progressive degradation of static backgrounds, and (ii) high-level semantic drift, such as inconsistent character identity and view-dependent attributes. To address this issue, EverAnimate restores drifted flow trajectories by anchoring generation to a persistent latent context memory, consisting of two complementary mechanisms. (i) Persistent Latent Propagation maintains a context memory across chunks to propagate identity and motion in latent space while mitigating temporal forgetting. (ii) Restorative Flow Matching introduces an implicit restoration objective during sampling through velocity adjustment, improving within-chunk fidelity. With only lightweight LoRA tuning, EverAnimate outperforms state-of-the-art long-animation methods in both short- and long-horizon settings: at 10 seconds, it improves PSNR/SSIM by 8%/7% and reduces LPIPS/FID by 22%/11%; at 90 seconds, the gains increase to 15%/15% and 32%/27%, respectively.

Understanding human motion beyond surface kinematics is crucial for motion analysis, rehabilitation, and injury risk assessment. However, progress in this domain is limited by the lack of large-scale datasets with biomechanical annotations, and by existing approaches that cannot directly infer internal biomechanical states from visual observations. In this paper, we introduce a simulation-based framework for estimating muscle activations from existing motion capture datasets, resulting in BioHuman10M, a large-scale dataset with synchronized video, motion, and activations. Building on BioHuman10M, we propose BioHuman, an end-to-end model that takes monocular video as input and jointly predicts human motion and muscle activations, effectively bridging visual observations and internal biomechanical states. Extensive experiments demonstrate that BioHuman enables accurate reconstruction of both kinematic motion and muscle activity, and generalizes across diverse subjects and motions. We believe our approach establishes a new benchmark for video-based biomechanical understanding and opens up new possibilities for physically grounded human modeling.

Trajectory-controlled human motion generation aims to synthesize realistic human motions conditioned on both textual descriptions and spatial trajectories. However, existing methods suffer from two critical limitations: first, the conflict between text and trajectory conditions disrupts the denoising process, resulting in compromised motion quality or inaccurate trajectory following; second, the use of redundant motion representations introduces inconsistencies between motion components, leading to instability during trajectory control. To address these challenges, we propose CMC, a decoupled framework that effectively coordinates text and trajectory conditions through a divide-and-conquer strategy. CMC follows a divide-and-conquer paradigm, comprising two cascaded stages: Trajectory Control and Motion Completion. In the first stage, a diffusion model generates a simplified representation of the controlled joints under trajectory guidance, based on the given trajectories, ensuring accurate and stable trajectory following. In the second stage, a text-conditioned diffusion inpainting model generates full-body motions using the simplified representation from the first stage as partial observations. To mitigate overfitting caused by limited inpainting training data, we further introduce the Selective Inpainting Mechanism (SIM), which alternates between text-to-motion generation and motion inpainting tasks during training. Experiments on HumanML3D and KIT datasets demonstrate that CMC achieves state-of-the-art performance in control accuracy and motion quality, demonstrating its effectiveness in coordinating multimodal conditions and representations.

The phrase 'human in the loop' is increasingly used to imply a sense of safety in relation to AI decision systems. It shouldn't. There are contexts where it can be applied appropriately, but these are not in the deployed decision systems we see dominating today. Human oversight of AI decision processes is one of the most popular proposals for addressing concerns, especially about bias, discrimination, misinformation, manipulation, accountability, and transparency. But there is insufficient examination of what human oversight actually means. The question raised in this paper is whether using the metaphor of a loop does anything to assist understanding of what is required and what is achieved in a particular decision context. Indiscriminate use of the loop metaphor obscures both processes and outcomes. It enables 'humanwashing', an activity analogous to 'greenwashing', where writers and commentators use language primarily aimed at putting systems in the best possible light.

This study investigates deep self-disclosure toward generative AI by examining perceived non-humanity and structural similarity as psychological factors beyond anthropomorphism. Perceived non-humanity may reduce evaluation apprehension, whereas structural similarity refers to the perceived logical alignment between a user's thinking and AI responses. Using cross-sectional survey data from 2,400 participants collected in 2025, this study analyzed associations with both the occurrence and depth of self-disclosure. Logistic regression indicated that the group high in both perceptions (Segment D) showed a significantly higher likelihood of disclosure than the baseline group (Segment A; OR = 11.35). ANOVA further showed significant between-group differences in disclosure depth. The findings suggest that trust-related behavior in deep self-disclosure may involve factors other than anthropomorphic perception. Because the study is exploratory and based on self-reported survey data, the results should be interpreted as associative rather than causal, and future longitudinal or experimental research is needed.

Personalisation is a standard feature of conversational AI systems used by millions; yet, the efficacy of personalisation methods is often evaluated in academic research using simulated users rather than real people. This raises questions about how users and their simulated counterparts differ in interaction patterns and judgements, as well as whether personalisation is best achieved through context-based prompting or weight-based fine-tuning. Here, in a large-scale within-subject experiment, we re-recruit 530 participants from 52 countries two years after they gave their preferences in the PRISM dataset (Kirk et al., 2024) to evaluate personalised and non-personalised language models in blinded multi-turn conversations. We find preference fine-tuning (P-DPO, Li et al., 2024) significantly outperforms both a generic model and personalised prompting but adapting to individual preference data yields marginal gains over training on pooled preferences from a diverse population. Beyond length biases, fine-tuning amplifies sycophancy and relationship-seeking behaviours that people reward in short-term evaluations but which may introduce deleterious long-term consequences. Replicating this within-subject experiment with simulated users recovers aggregate model hierarchies but simulators perform far below human self-consistency baselines for individual judgements, discuss different topics, exhibit amplified position biases, and produce feedback dynamics that diverge from humans.

Evaluating whether large vision-language models (VLMs) align with human perception for high-level semantic scene comprehension remains a challenge. Traditional white-box interpretability methods are inapplicable to closed-source architectures and passive metrics fail to isolate causal features. We introduce Counterfactual Semantic Saliency (CSS). This black-box, model-agnostic framework quantifies the importance of objects by measuring the semantic shift induced by their causal ablation from a scene. To evaluate AI-human semantic alignment, we tested prominent VLMs against a human psychophysics baseline comprising 16,289 valid responses across 307 complex natural scenes and 1,306 high-fidelity counterfactual variants. Our analysis reveals a pervasive scene comprehension gap: models exhibit an overreliance (relative to humans) on large objects (size bias), objects at the center of the image (center bias), and high saliency objects. In contrast, models rely less on people in the scenes than our human participants to describe the images. A model's size bias is a primary driver explaining variations in model-human semantic divergence. Code and data will be available at https://github.com/starsky77/Counterfactual-Semantic-Saliency.

Emotion expression is central to human--robot interaction, yet little is known about how people interpret affect on robots with sparse, non-anthropomorphic expressive capabilities. This study examined how people perceive emotional expressions displayed by Reachy Mini (Pollen Robotics and Hugging Face), a low-degree-of-freedom (low-DoF) robot with a constrained and distinctly non-human expressive repertoire. In an online within-subjects study, 100 participants viewed 10 short video clips of Reachy Mini expressing different emotions and, for each clip, identified the perceived emotion, rated its valence and arousal, and evaluated the robot on social-perception traits. Exact emotion recognition was modest overall and varied considerably across expressions, with anger, sadness, and interest recognized more reliably than emotions such as love, pleasure, shame, and disgust. However, participants were generally more successful at recovering broader affective meaning than exact emotion labels, particularly along valence and arousal dimensions. Emotional expressions also shaped social evaluation, as positive expressions were perceived as warmer and more sociable than negative ones, and animacy varied less across conditions. These findings suggest that even constrained robotic expressions can communicate affective meaning and influence social impressions, positioning Reachy Mini as a useful benchmark for studying affective communication in low-DoF robots.

High-quality labeled data is essential for training robust machine learning models, yet obtaining annotations at scale remains expensive. AI-assisted annotation has therefore become standard in large-scale labeling workflows. However, in tasks where model predictions carry two independent components, a class label and spatial boundaries, a model may classify an object with high confidence while mislocalizing it. Existing AI-assisted workflows offer annotators no signal about where spatial errors are most likely. Without such guidance, humans may systematically underinspect subtly misplaced boxes. We address this by studying the effect of visualizing spatial uncertainty via a purpose-built interface. In a controlled study with 120 participants, those receiving uncertainty cues achieve higher label quality while being faster overall. A box-level analysis confirms that the cues redirect annotator effort toward high-uncertainty predictions and away from well-localized boxes. These findings establish localization uncertainty as a lever to improve human-in-the-loop annotation. Code is available at https://mos-ks.github.io/MUHA/.

Generative Artificial Intelligence (AI) tools are rapidly adopted in the workplace and in education, yet the empirical evidence on AI's impact remains mixed. We propose a model of human-AI interaction to better understand and analyze several mechanisms by which AI affects productivity. In our setup, human agents with varying skill levels exert utility-maximizing effort to produce certain task outcomes with AI assistance. We find that incorporating either endogeneity in skill development or in AI unreliability can induce a productivity paradox: increased levels of AI assistance may degrade productivity, leading to potentially significant shortfalls. Moreover, we examine the long-term distributional effect of AI on skill, and demonstrate that skill polarization can emerge in steady state when accounting for heterogeneity in AI literacy -- the agent's capability to identify and adapt to inaccurate AI outputs. Our results elucidate several mechanisms that may explain the emergence of human-AI productivity paradoxes and skill polarization, and identify simple measures that characterize when they arise.

As AI agents transition from research prototypes to enterprise production systems, the tool interfaces they consume remain rooted in human-oriented CRUD paradigms. This paper identifies five fundamental architectural mismatches between conventional APIs and autonomous agent requirements: exact-identifier dependence, rendering-oriented responses, single-shot interaction assumptions, user-equivalent authorization, and opaque error semantics. We propose the Agent-First Tool API paradigm, comprising three integrated mechanisms: (1) a Six-Verb Semantic Protocol that decomposes tool interactions into search, resolve, preview, execute, verify, and recover phases; (2) a Normalized Tool Contract (NTC) providing structured decision-support metadata including confidence scores, evidence chains, and suggested next actions; and (3) a dual-layer governance pipeline combining static capability policies with dynamic risk escalation. The paradigm is implemented and validated in a production multi-tenant SaaS platform serving 85 registered tools across 6 business domains. Comparative experiments on 50 real operational tasks demonstrate that Agent-First APIs achieve 88% end-to-end task success rate versus 64% for optimized CRUD baselines (+37.5%), while reducing required human interventions by 72.7% and improving autonomous error recovery by 5.8x. We establish that the paradigm is orthogonal and complementary to transport-layer standards such as MCP, operating as the semantic application layer above existing tool discovery and invocation protocols.

A core challenge in program synthesis is online library learning: the incremental acquisition of reusable abstractions under uncertainty about future task demands. Existing algorithms treat library learning as retrospective compression over a static task distribution, where the learned library is determined by the corpus of past tasks. However, real-world learning domains are often non-stationary, with tasks arising from a generative process that evolves over time. We propose and test the hypothesis that in non-stationary domains human library learning selects abstractions prospectively: targeting compression of future tasks. We study this question using the Pattern Builder Task, a visual program synthesis paradigm in which participants construct increasingly complex geometric patterns from a small set of primitives, transformations, and custom helpers that carry forward across trials. Using this task, we conduct two experiments with complementary latent curricula, designed to dissociate between behaviors consistent with prospective compression, and alternative library learning accounts. Using six computational models spanning online library learning strategies, we show that human abstraction behavior reflects sensitivity to latent, non-stationary structure in the task-generating process. This behavior is consistent with prospective compression, and cannot be captured by existing retrospective compression-based algorithms, or inductive biases modeled by LLM-based program synthesis.

Circuit discovery aims to explain how language models (LMs) implement a specific task by localizing and interpreting a circuit, a computational subgraph responsible for the LM's behavior. Existing circuit discovery methods are hypothesis-driven; they first informally define a task with a dataset, and then apply a circuit discovery algorithm over that dataset to obtain a single circuit. This imposes two strong assumptions: that the LM implements the task with a single circuit, and that the dataset adequately represents the task as humans understand it. We systematically test these assumptions across four previously studied tasks and find that even minor dataset variations that preserve task semantics can produce circuits with low edge overlap and cross-dataset faithfulness. More strikingly, when applied to a mixed dataset with two distinct tasks whose separately discovered circuits have near-zero cross-faithfulness, existing methods still return a single circuit with high faithfulness across both tasks. This indicates that current methods discover dataset-specific circuits, rather than general task circuits. We propose Data-driven Circuit Discovery (DCD), a new discovery framework that drops both assumptions: instead of returning a single circuit for a dataset, DCD first clusters examples in the dataset by how similarly the model processes them and discovers a separate circuit for each group. This allows distinct mechanisms to appear separately rather than merged into a single circuit; each circuit explains its group, not the full task. Experiments show that DCD discovers multiple circuits per dataset, each more faithful to its group than a single circuit discovered by existing methods. Broadly, DCD lets the data reveal mechanistic structure within LMs, rather than relying on human-defined task boundaries that may not align with how models organize their computation.

Abstract concepts - justice, theory, availability - have no single perceivable referent; in the human brain, their meaning emerges from a web of experiences, affect, and social context. Do large language models (LLMs) ground abstract concepts in a similar way? We study this by replicating property-generation experiments from cognitive science on 21 frontier and open-weight LLMs. Across models and experiments, we find a consistent pattern: when compared to humans, models rely too heavily on word associations, and underproduce properties tied to emotion and internal states. This yields a large and consistent grounding gap: no model exceeds a Pearson correlation r=0.37 with human responses, compared to a human-to-human ceiling above r=0.9. To better interpret this gap, we also replicate a rating experiment on grounding categories and find that here LLMs align more closely with human judgment, and alignment improves as models get larger. We then use sparse autoencoders (SAEs) to inspect whether this information is also reflected in the models' internal features, and we do identify features connected to grounding dimensions such as "sensorimotor" and "social". These findings suggest that current LLMs can recover grounding dimensions when explicitly queried, but do not recruit them in a human-like way when words are generated freely.