Generative AI Evaluation

Scientific peer review faces mounting strain as submission volumes surge, making it increasingly difficult to sustain review quality, consistency, and timeliness. Recent advances in AI have led the community to consider its use in peer review, yet a key unresolved question is whether AI can generate technically sound reviews at real-world conference scale. Here we report the first large-scale field deployment of AI-assisted peer review: every main-track submission at AAAI-26 received one clearly identified AI review from a state-of-the-art system. The system combined frontier models, tool use, and safeguards in a multi-stage process to generate reviews for all 22,977 full-review papers in less than a day. A large-scale survey of AAAI-26 authors and program committee members showed that participants not only found AI reviews useful, but actually preferred them to human reviews on key dimensions such as technical accuracy and research suggestions. We also introduce a novel benchmark and find that our system substantially outperforms a simple LLM-generated review baseline at detecting a variety of scientific weaknesses. Together, these results show that state-of-the-art AI methods can already make meaningful contributions to scientific peer review at conference scale, opening a path toward the next generation of synergistic human-AI teaming for evaluating research.

While recent research has explored AI tools' ability to classify the quality of mathematical tasks (arXiv:2603.03512), little is known about their capacity to increase the quality of existing tasks. This study investigated whether AI tools could successfully upgrade low-cognitive-demand mathematics tasks. Eleven tools were tested, including six broadly available, general-purpose AI tools (e.g., ChatGPT and Claude) and five tools specialized for mathematics teachers (e.g., Khanmigo, coteach.ai). Using the Task Analysis Guide framework (Stein & Smith, 1998), we prompted AI tools to modify two different types of low-demand mathematical tasks. The prompting strategy aimed to represent likely approaches taken by knowledgeable teachers, rather than extensive optimization to find a more effective prompt (i.e., an optimistic typical outcome). On average, AI tools were only moderately successful: tasks were accurately upgraded only 64% of the time, with different AI tool performance ranging from quite weak (33%) to broadly successful (88%). Specialized tools were only moderately more successful than general-purpose tools. Failure modes included both "undershooting" (maintaining low cognitive demand) and "overshooting" (elevating tasks to an overly ambitious target category that likely would be rejected by teachers). Interestingly, there was a small negative correlation (r = -.35) between whether a given AI tool was able to correctly classify the cognitive demand of tasks and whether the AI was able to upgrade tasks, showing that the ability to modify tasks (i.e., a generative task) represents a distinct capability from the ability to classify them (i.e., judgement using a rubric). These findings have important implications for understanding AI's potential role in curriculum adaptation and highlight the need for specialized approaches to support teachers in modifying instructional materials.

Identifying promising research directions in fast-moving subareas is one of the most cognitively expensive tasks in modern AI research. Existing LLM-driven scientific discovery systems are typically limited to one-shot prompting on static literature snapshots and are validated only against contemporary judges such as human reviewers, agent peer review, wet-lab assays, or self-evaluation, leaving open whether they can anticipate future trends. We present Continuous Knowledge Metabolism (CKM), an AI workflow for hypothesis generation with three key capabilities: (i) continuous literature metabolism via sliding windows that maintain an evolving knowledge state; (ii) predictive evaluation, which grades hypotheses against papers published after the generation window; and (iii) practitioner-grade failure detection that diagnoses workflow failure modes from its outputs. On a 50-topic machine learning benchmark, CKM-Lite produces at least one validated hypothesis on 72% of topics (36 out of 50), more than doubling a one-shot baseline (30%) at approximately 3 dollars per topic and achieving 91% lower token cost. Validated hypotheses precede their matched papers by an average of 404 days (55 hits across 36 topics; median 399 days, range 66-757 days). Broadly, predictive validation against future literature provides a falsifiable, low-cost alternative to contemporary-judge evaluation protocols and can be applied wherever a corpus has dated publication records.

Welcome to the ninth edition of the AI Index report. As AI continues to advance rapidly, the question becomes whether the systems built around it can keep up. Governance frameworks, evaluation methods, education systems, and the data infrastructure needed to track AI's impact are struggling to match the pace of the technology itself. That gap between what AI can do and how prepared we are to manage it runs through every chapter of this year's report. New in this edition, the report tracks how AI is being tested more ambitiously across reasoning, safety, and real-world task execution, and why those measurements are increasingly difficult to rely on. It also features new estimates of generative AI's economic value alongside emerging evidence of its labor market effects, an analytical framework on AI sovereignty, and a science chapter developed in collaboration with Schmidt Sciences. For the first time, the report features standalone chapters on AI in science and AI in medicine, reflecting AI's growing impact across these two domains.

We present the progress of the GPT family from GPT-3 through GPT-3.5, GPT-4, GPT-4 Turbo, GPT-4o, GPT-4.1, and the GPT-5 family. Our work is comparative rather than merely historical. We investigates how the family evolved in technical framing, user interaction, modality, deployment architecture, and governance viewpoint. The work focuses on five recurring themes: technical progression, capability changes, deployment shifts, persistent limitations, and downstream consequences. In term of research design, we consider official technical reports, system cards, API and model documentation, product announcements, release notes, and peer-reviewed secondary studies. A primary assertion is that later GPT generations should not be interpreted only as larger or more accurate language models. Instead, the family evolves from a scaled few-shot text predictor into a set of aligned, multimodal, tool-oriented, long-context, and increasingly workflow-integrated systems. This development complicates simple model-to-model comparison because product routing, tool access, safety tuning, and interface design become part of the effective system. Across generations, several limitations remain unchanged: hallucination, prompt sensitivity, benchmark fragility, uneven behavior across domains and populations, and incomplete public transparency about architecture and training. However, the family has evolved software development, educational practice, information work, interface design, and discussions of frontier-model governance. We infer that the transition from GPT-3 to GPT-5 is best understood not only as an improvement in model capability, but also as a broader reformulation of what a deployable AI system is, how it is evaluated, and where responsibility should be located when such systems are used at scale.

In the psychological literature the term general intelligence' describes correlations between abilities and not simply the number of abilities. This paper connects Spearman's $g$-factor from psychometrics, measuring a positive manifold, to the implicit ``$G$-factor'' in claims about artificial general intelligence (AGI) performance on temporally structured benchmarks. By treating LLM benchmark batteries as cognitive test batteries and model releases as subjects, principal component analysis is applied to a models $\times$ benchmarks $\times$ time matrix spanning 39 models (2019--2025) and 14 benchmarks. Preliminary results confirm a strong positive manifold in which all 28 pairwise correlations positive across 8 benchmarks. By analyzing the spectrum of the benchmark correlation through time, PC1 explains 90\% of variance on a 5-benchmark core battery ($n=19$)) reducing to 77\% by 2024. On a four benchmark battery, PC1 is found to peak at 92\% of the variance between 2023--2024 and reduce to 64\% with the arrival of reasoning-specialized models in 2024. This is coincident with a rotation in the G-factor as models outsource reasoning' to tools. The analysis of partial correlation matrices through time provides evidence for the evolution of specialization beneath the positive manifold of general intelligence (AI-hedgehog) encompassing diverse high dimensional problem solving systems (AI-foxes). In strictly psychometric terms, AI models exhibit general intelligence suppressing specialized intelligences. LLMs invert the ideal of substituting complicated models with parsimonious mechanisms, a `Ptolemaic Succession' of theories, with architectures of increasing hierarchical complication and capability.

Accurate evaluation is central to the large language model (LLM) ecosystem, guiding model selection and downstream adoption across diverse use cases. In practice, however, evaluating generative outputs typically relies on rigid lexical methods to extract and assess answers, which can conflate a model's true problem-solving ability with its compliance with predefined formatting guidelines. While recent LLM-as-a-Judge approaches mitigate this issue by assessing semantic correctness rather than strict structural conformity, they also introduce substantial computational overhead, making evaluation costly. In this work, we first systematically investigate the limitations of lexical evaluation through a large-scale empirical study spanning 36 models and 15 downstream tasks, demonstrating that such methods correlate poorly with human judgments. To address this limitation, we introduce BERT-as-a-Judge, an encoder-driven approach for assessing answer correctness in reference-based generative settings, robust to variations in output phrasing, and requiring only lightweight training on synthetically annotated question-candidate-reference triplets. We show that it consistently outperforms the lexical baseline while matching the performance of much larger LLM judges, providing a compelling trade-off between the two and enabling reliable, scalable evaluation. Finally, through extensive experimentation, we provide detailed insights into BERT-as-a-Judge's performance to offer practical guidance for practitioners, and release all project artifacts to foster downstream adoption.

Generative AI is rapidly transforming how organizations create value and evaluate talent. While large language models enhance baseline output quality, they simultaneously introduce ambiguity in assessing human creativity, as observable artifacts may be partially or fully AI-generated. This paper reconceptualizes creativity as a distributional and process-based property that emerges under shared constraints and competitive incentives. We introduce a quantitative framework for measuring creativity as novelty in synthesis, operationalized through idea generation and idea transformation within embedding space. Empirical evaluation demonstrates that the proposed metrics align with intuitive judgments of creativity while capturing distinctions that surface-level quality assessments miss. We further identify a structural shift toward bimodal distributions of creative output in AI-mediated environments, with implications for hiring, leadership, and competitive strategy. The findings suggest that in the age of generative AI, distinctiveness rather than fluency becomes the primary signal of human creative capability.

Text-to-Audio-Video (T2AV) generation is rapidly becoming a core interface for media creation, yet its evaluation remains fragmented. Existing benchmarks largely assess audio and video in isolation or rely on coarse embedding similarity, failing to capture the fine-grained joint correctness required by realistic prompts. We introduce AVGen-Bench, a task-driven benchmark for T2AV generation featuring high-quality prompts across 11 real-world categories. To support comprehensive assessment, we propose a multi-granular evaluation framework that combines lightweight specialist models with Multimodal Large Language Models (MLLMs), enabling evaluation from perceptual quality to fine-grained semantic controllability. Our evaluation reveals a pronounced gap between strong audio-visual aesthetics and weak semantic reliability, including persistent failures in text rendering, speech coherence, physical reasoning, and a universal breakdown in musical pitch control. Code and benchmark resources are available at http://aka.ms/avgenbench.

As large language model-based chat systems become increasingly widely used, generative engine optimization (GEO) has emerged as an important problem for information access and retrieval. In classical search engines, results are comparatively transparent and stable: a single query often provides a representative snapshot of where a page or brand appears relative to competitors. The inherent probabilistic nature of AI search changes this paradigm. Answers can vary across runs, prompts, and time, making one-off observations unreliable. Drawing on empirical studies, our findings underscore the need for repeated measurements to assess a brand's GEO performance and to characterize visibility as a distribution rather than a single-point outcome.

Statistical evaluation aims to estimate the generalization performance of a model using held-out i.i.d. test data sampled from the ground-truth distribution. In supervised learning settings such as classification, performance metrics such as error rate are well-defined, and test error reliably approximates population error given sufficiently large datasets. In contrast, evaluation is more challenging for generative models due to their open-ended nature: it is unclear which metrics are appropriate and whether such metrics can be reliably evaluated from finite samples. In this work, we introduce a theoretical framework for evaluating generative models and establish evaluability results for commonly used metrics. We study two categories of metrics: test-based metrics, including integral probability metrics (IPMs), and Rényi divergences. We show that IPMs with respect to any bounded test class can be evaluated from finite samples up to multiplicative and additive approximation errors. Moreover, when the test class has finite fat-shattering dimension, IPMs can be evaluated with arbitrary precision. In contrast, Rényi and KL divergences are not evaluable from finite samples, as their values can be critically determined by rare events. We also analyze the potential and limitations of perplexity as an evaluation method.

The capabilities of artificial intelligence (AI) lie along a jagged frontier, where AI systems surprisingly fail on tasks that humans find easy and succeed on tasks that humans find hard. To investigate user reactions to this phenomenon, we developed an incentive-compatible experimental methodology based on diagram generation tasks, in which we induce errors in generative AI output and test effects on user reliance. We demonstrate the interface in a preregistered 3x2 experiment (N = 577) with error rates of 10%, 30%, or 50% on easier or harder diagram generation tasks. We confirmed that observing more errors reduces use, but we unexpectedly found that easy-task errors did not significantly reduce use more than hard-task errors, suggesting that people are not averse to jaggedness in this experimental setting. We encourage future work that varies task difficulty at the same time as other features of AI errors, such as whether the jagged error patterns are easily learned.

Large Language Models (LLMs) can compose poetry, but how far are they from human poets? In this paper, we introduce POEMetric, the first comprehensive framework for poetry evaluation, examining 1) basic instruction-following abilities in generating poems according to a certain form and theme, 2) advanced abilities of showing creativity, lexical diversity, and idiosyncrasy, evoking emotional resonance, and using imagery and literary devices, and 3) general appraisal of the overall poem quality and estimation of authorship. We curated a human poem dataset - 203 English poems of 7 fixed forms annotated with meter, rhyme patterns and themes - and experimented with 30 LLMs for poetry generation based on the same forms and themes of the human data, totaling 6,090 LLM poems. Based on POEMetric, we assessed the performance of both human poets and LLMs through rule-based evaluation and LLM-as-a-judge, whose results were validated by human experts. Results show that, though the top model achieved high form accuracy (4.26 out of 5.00, with Gemini-2.5-Pro as a judge; same below) and theme alignment (4.99), all models failed to reach the same level of advanced abilities as human poets, who achieved unparalleled creativity (4.02), idiosyncrasy (3.95), emotional resonance (4.06), and skillful use of imagery (4.49) and literary devices (4.67). Humans also defeated the best-performing LLM in overall poem quality (4.22 vs. 3.20). As such, poetry generation remains a formidable challenge for LLMs. Data and codes are released at https://github.com/Bingru-Li/POEMetric.

Autonomous AI systems can now generate complete economics research papers, but they substantially underperform human-authored publications in head-to-head comparisons. This paper decomposes the quality gap into two independent components: research idea quality and execution quality. Using a two-model ensemble of fine-tuned language models trained on publication decisions (Gong, Li, and Zhou, 2026) to evaluate idea quality and a comprehensive six-dimension rubric assessed by Gemini 3.1 Flash Lite -- the same model family used as the APE tournament judge, ensuring methodological consistency -- to evaluate execution quality, we analyze 953 economics papers -- 912 AI-generated papers from the APE project and 41 human papers published in the American Economic Review and AEJ: Economic Policy. The idea quality gap is large (Cohen's d = 2.23, p < 0.001), with human papers achieving 47.1% mean ensemble exceptional probability versus 16.5% for AI. The execution quality gap is also significant but smaller (d = 0.90, p < 0.001), with human papers scoring 4.38/5.0 versus 3.84. Idea quality accounts for approximately 71% of the overall quality difference, with execution contributing 29%. The largest execution weakness is mechanism analysis depth (d = 1.43); no significant difference is found on robustness. We document that 74% of AI papers employ difference-in-differences, and only 7 AI papers (0.8%) surpass the median human paper on both idea and execution quality simultaneously. The primary bottleneck to competitive AI-generated economics research remains ideation.

Diffusion language models have seen exciting recent progress, offering far more flexibility in generative trajectories than autoregressive models. This flexibility has motivated a growing body of research into new approaches to diffusion language modeling, which typically begins at the scale of GPT-2 small (150 million parameters). However, these advances introduce new issues with evaluation methodology. In this technical note, we discuss the limitations of current methodology and propose principled augmentations to ensure reliable comparisons. We first discuss why OpenWebText has become the standard benchmark, and why alternatives such as LM1B are inherently less meaningful. We then discuss the limitations of likelihood evaluations for diffusion models, and explain why relying on generative perplexity alone as a metric can lead to uninformative results. To address this, we show that generative perplexity and entropy are two components of the KL divergence to a reference distribution. This decomposition explains generative perplexity's sensitivity to entropy, and naturally suggests generative frontiers as a principled method for evaluating model generative quality. We conclude with empirical observations on model quality at this scale. We include a blog post with interactive content to illustrate the argument at https://patrickpynadath1.github.io/blog/eval_methodology/.

We characterize AI automation as a continuum between crashing waves, in which capabilities jump abruptly across narrow task sets, and rising tides, in which capabilities improve continuously and broadly. Using evidence from more than 6,000 text-based, LLM-addressable tasks derived from the U.S. Department of Labor's O*NET taxonomy and over 60,000 evaluations by experienced workers, we find little evidence of crashing waves (contrary to existing views). Instead, rising tides are the primary form of AI progress. AI performance is high and improving rapidly across many tasks. In 2024-Q2, models completed text-based tasks that take humans about 1.5 hours to complete with roughly 60% success, rising above 70% by 2025-Q3. If recent trends in AI capability growth persist, frontier LLMs will be able to complete most text-based tasks at minimally sufficient quality with 88%-97% success by 2030.

This paper introduces the first systematic evaluation framework for quantifying the quality and risks of papers written by modern coding agents. While AI-driven paper writing has become a growing concern, rigorous evaluation of the quality and potential risks of AI-written papers remains limited, and a unified understanding of their reliability is still lacking. We introduce Paper Reconstruction Evaluation (PaperRecon), an evaluation framework in which an overview (overview.md) is created from an existing paper, after which an agent generates a full paper based on the overview and minimal additional resources, and the result is subsequently compared against the original paper. PaperRecon disentangles the evaluation of the AI-written papers into two orthogonal dimensions, Presentation and Hallucination, where Presentation is evaluated using a rubric and Hallucination is assessed via agentic evaluation grounded in the original paper source. For evaluation, we introduce PaperWrite-Bench, a benchmark of 51 papers from top-tier venues across diverse domains published after 2025. Our experiments reveal a clear trade-off: while both ClaudeCode and Codex improve with model advances, ClaudeCode achieves higher presentation quality at the cost of more than 10 hallucinations per paper on average, whereas Codex produces fewer hallucinations but lower presentation quality. This work takes a first step toward establishing evaluation frameworks for AI-driven paper writing and improving the understanding of its risks within the research community.

Evaluating humor in large language models (LLMs) is an open challenge because existing approaches yield isolated, incomparable metrics rather than unified model rankings, making it difficult to track progress across systems. We introduce HumorRank, a tournament-based evaluation framework and leaderboard for textual humor generation. On two public benchmarks (SemEval-2026 MWAHAHA and Humor Transfer Bench), we conduct extensive automated pairwise evaluation across nine models spanning proprietary, open-weight, and specialized systems. Pairwise judgments are produced by LLM judges grounded in the General Theory of Verbal Humor (GTVH): each judge integrates structured comedic analysis into adjudication, jointly yielding a preference decision, an interpretable rationale, and mechanism, delivery, and failure tags rather than a black-box funniness score. Judgments are aggregated via an Adaptive Swiss tournament, with Bradley-Terry Maximum Likelihood Estimation (MLE) producing globally consistent humor generation capability rankings. Rankings are cross-judge stable: independent LLM judges (Llama 3.3 70B and Qwen 2.5 72B) yield Kendall tau = 0.889 on both benchmarks, and a human calibration study shows human-LLM agreement tracking human-human agreement on hard funny-versus-funny pairs. Our results demonstrate that HumorRank yields statistically grounded model stratifications, showing that humor quality is associated with mastery of comedic mechanisms such as incongruity, conciseness, escalation, and absurdity rather than model scale alone, with specialized fine-tuned models reaching parity with far larger systems. HumorRank thus provides a scalable, interpretable, and reproducible methodology for benchmarking and understanding LLM-generated humor.

Software engineering courses often require rapid upskilling in supporting knowledge areas such as domain understanding and modeling methods. We report an experience from a two-week milestone in a master's course where 29 students used a customized ChatGPT (GPT-3.5) tutor grounded in a curated course knowledge base to learn cryptocurrency-finance basics and Domain-Driven Design (DDD). We logged all interactions and evaluated a 34.5% random sample of prompt-answer pairs (60/~174) with a five-dimension rubric (accuracy, relevance, pedagogical value, cognitive load, supportiveness), and we collected pre/post self-efficacy. Responses were consistently accurate and relevant in this setting: accuracy averaged 98.9% with no factual errors and only 2/60 minor inaccuracies, and relevance averaged 92.2%. Pedagogical value was high (89.4%) with generally appropriate cognitive load (82.78%), but supportiveness was low (37.78%). Students reported large pre-post self-efficacy gains for genAI-assisted domain learning and DDD application. From these observations we distill seventeen concrete teaching practices spanning prompt/configuration and course/workflow design (e.g., setting expected granularity, constraining verbosity, curating guardrail examples, adding small credit with a simple quality rubric). Within this single-course context, results suggest that genAI-supported learning can complement instruction in domain understanding and modeling tasks, while leaving room to improve tone and follow-up structure.

Generative AI systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be measured rather than fundamental to the system's operation. Drawing on hermeneutic theory from the humanities, we argue that GenAI systems function as "context machines" that must inherently address three interpretive challenges: situatedness (meaning only emerges in context), plurality (multiple valid interpretations coexist), and ambiguity (interpretations naturally conflict). We present computational hermeneutics as an emerging framework offering an interpretive account of what GenAI systems do, and how they might do it better. We offer three principles for hermeneutic evaluation -- that benchmarks should be iterative, not one-off; include people, not just machines; and measure cultural context, not just model output. This perspective offers a nascent paradigm for designing and evaluating contemporary AI systems: shifting from standardized questions about accuracy to contextual ones about meaning.

AI arenas, which rank generative models from pairwise preferences of users, are a popular method for measuring the relative performance of models in the course of their organic use. Because rankings are computed from noisy preferences, there is a concern that model producers can exploit this randomness by submitting many models (e.g., multiple variants of essentially the same model) and thereby artificially improve the rank of their top models. This can lead to degradations in the quality, and therefore the usefulness, of the ranking. In this paper, we begin by establishing, both theoretically and in simulations calibrated to data from the platform Arena (formerly LMArena, Chatbot Arena), conditions under which producers can benefit from submitting clones when their goal is to be ranked highly. We then propose a new mechanism for ranking models from pairwise comparisons, called You-Rank-We-Rank (YRWR). It requires that producers submit rankings over their own models and uses these rankings to correct statistical estimates of model quality. We prove that this mechanism is approximately clone-robust, in the sense that a producer cannot improve their rank much by doing anything other than submitting each of their unique models exactly once. Moreover, to the extent that model producers are able to correctly rank their own models, YRWR improves overall ranking accuracy. In further simulations, we show that indeed the mechanism is approximately clone-robust and quantify improvements to ranking accuracy, even under producer misranking.

Converting a pretrained Transformer into a more efficient hybrid model through distillation offers a promising approach to reducing inference costs. However, achieving high-quality generation in distilled models requires careful joint design of both the student architecture and the distillation process. Many prior distillation works evaluate downstream multiple-choice benchmarks by ranking candidate answers with log-likelihood rather than requiring autoregressive generation, which can obscure important differences in model quality. For example, on overlapping benchmarks, we show that a 7B distilled model that nearly matches its teacher to within 0.2 pp under log-likelihood scoring falls behind by 20.8 pp when it must generate answers autoregressively. We investigate this phenomenon with GenDistill, a multi-stage pipeline we designed for distilling a pretrained Transformer into an efficient Hybrid Kimi Delta Attention (Hybrid-KDA) student. Using it as a controlled testbed on Qwen3-0.6B, we systematically ablate six design axes (training objective, loss masking, training duration, dataset selection, parameter freezing, and architecture choice) and evaluate every choice under both log-likelihood and generation-based protocols. We find that log-likelihood-based evaluation consistently underestimates the gap between teacher and student, and can in some cases reverse the ranking of design choices, so conclusions drawn from perplexity-only evaluation may be misleading. Among the factors we study, dataset selection, completion-only masking, and freezing attention layers during post-training have the largest impact on generation quality. Our best distillation recipe, using a Hybrid-KDA model as the student, retains 86-90% of teacher accuracy on knowledge benchmarks while reducing KV cache memory by up to 75% and improving time-to-first-token by 2-4x at 128K-token contexts.

Generative modeling has become a central paradigm in protein research, extending machine learning beyond structure prediction toward sequence design, backbone generation, inverse folding, and biomolecular interaction modeling. However, the literature remains fragmented across representations, model classes, and task formulations, making it difficult to compare methods or identify appropriate evaluation standards. This survey provides a systematic synthesis of generative AI in protein research, organized around (i) foundational representations spanning sequence, geometric, and multimodal encodings; (ii) generative architectures including $\mathrm{SE}(3)$-equivariant diffusion, flow matching, and hybrid predictor-generator systems; and (iii) task settings from structure prediction and de novo design to protein-ligand and protein-protein interactions. Beyond cataloging methods, we compare assumptions, conditioning mechanisms, and controllability, and we synthesize evaluation best practices that emphasize leakage-aware splits, physical validity checks, and function-oriented benchmarks. We conclude with critical open challenges: modeling conformational dynamics and intrinsically disordered regions, scaling to large assemblies while maintaining efficiency, and developing robust safety frameworks for dual-use biosecurity risks. By unifying architectural advances with practical evaluation standards and responsible development considerations, this survey aims to accelerate the transition from predictive modeling to reliable, function-driven protein engineering.

Beneath the stunning visual fidelity of modern AIGC models lies a "logical desert", where systems fail tasks that require physical, causal, or complex spatial reasoning. Current evaluations largely rely on superficial metrics or fragmented benchmarks, creating a performance mirage'' that overlooks the generative process. To address this, we introduce ViGoR Vision-G}nerative Reasoning-centric Benchmark), a unified framework designed to dismantle this mirage. ViGoR distinguishes itself through four key innovations: 1) holistic cross-modal coverage bridging Image-to-Image and Video tasks; 2) a dual-track mechanism evaluating both intermediate processes and final results; 3) an evidence-grounded automated judge ensuring high human alignment; and 4) granular diagnostic analysis that decomposes performance into fine-grained cognitive dimensions. Experiments on over 20 leading models reveal that even state-of-the-art systems harbor significant reasoning deficits, establishing ViGoR as a critical stress test'' for the next generation of intelligent vision models. The demo have been available at https://vincenthancoder.github.io/ViGoR-Bench/

Evaluating the pedagogical quality of AI tutors remains challenging: standard NLG metrics do not determine whether responses identify mistakes, scaffold reasoning, or avoid revealing the answers. For the task of mistake remediation, we derive a hierarchy of pedagogical aspects from human pairwise preferences on MRBench, and synthesize minimally contrastive response pairs that differ along key aspects (e.g., mistake identification and location, targetedness, scaffolding, actionability, clarity, and coherence). We develop and release Bradley-Terry preference models trained on weighted-sum rankings that we automatically create from MRBench, synthetic pairs, and data combinations. Using only synthetic data, our best model reaches 0.69 pairwise accuracy on a human preference test, and combining weighted-sum data with targeted synthetic groups improves accuracy to 0.74, outperforming larger general-purpose reward models while using only a 0.5B-parameter backbone.

The rapid adoption of large language models (LLMs) in education raises profound challenges for assessment design. To adapt assessments to the presence of LLM-based tools, it is crucial to characterize the strengths and weaknesses of LLMs in a generalizable, valid and reliable manner. However, current LLM evaluations often rely on descriptive statistics derived from benchmarks, and little research applies theory-grounded measurement methods to characterize LLM capabilities relative to human learners in ways that directly support assessment design. Here, by combining educational data mining and psychometric theory, we introduce a statistically principled approach for identifying items on which humans and LLMs show systematic response differences, pinpointing where assessments may be most vulnerable to AI misuse, and which task dimensions make problems particularly easy or difficult for generative AI. The method is based on Differential Item Functioning (DIF) analysis -- traditionally used to detect bias across demographic groups -- together with negative control analysis and item-total correlation discrimination analysis. It is evaluated on responses from human learners and six leading chatbots (ChatGPT-4o & 5.2, Gemini 1.5 & 3 Pro, Claude 3.5 & 4.5 Sonnet) to two instruments: a high school chemistry diagnostic test and a university entrance exam. Subject-matter experts then analyzed DIF-flagged items to characterize task dimensions associated with chatbot over- or under-performance. Results show that DIF-informed analytics provide a robust framework for understanding where LLM and human capabilities diverge, and highlight their value for improving the design of valid, reliable, and fair assessment in the AI era.

Architecture views are essential for software architecture documentation, yet their manual creation is labor intensive and often leads to outdated artifacts. As systems grow in complexity, the automated generation of views from source code becomes increasingly valuable. Goal: We empirically evaluate the ability of LLMs and agentic approaches to generate architecture views from source code. Method: We analyze 340 open-source repositories across 13 experimental configurations using 3 LLMs with 3 prompting techniques and 2 agentic approaches, yielding 4,137 generated views. We evaluate the generated views by comparing them with the ground-truth using a combination of automated metrics complemented by human evaluations. Results: Prompting strategies offer marginal improvements. Few-shot prompting reduces clarity failures by 9.2% compared to zero-shot baselines. The custom agentic approach consistently outperforms the general-purpose agent, achieving the best clarity (22.6% failure rate) and level-of-detail success (50%). Conclusions: LLM and agentic approaches demonstrate capabilities in generating syntactically valid architecture views. However, they consistently exhibit granularity mismatches, operating at the code level rather than architectural abstractions. This suggests that there is still a need for human expertise, positioning LLMs and agents as assistive tools rather than autonomous architects.

The ubiquity of multimedia content is reshaping online information spaces, particularly in social media environments. At the same time, search is being rapidly transformed by generative AI, with large language models (LLMs) routinely deployed as intermediaries between users and multimedia content to retrieve and summarize information. Despite their growing influence, the impact of LLM inaccuracies and potential vulnerabilities on multimedia information-seeking tasks remains largely unexplored. We investigate how generative AI affects accuracy, efficiency, and confidence in information retrieval from videos. We conduct an experiment with around 900 participants on 8,000+ video-based information-seeking tasks, comparing behavior across three conditions: (1) access to videos only, (2) access to videos with LLM-based AI assistance, and (3) access to videos with a deceiving AI assistant designed to provide false answers. We find that AI assistance increases accuracy by 3-7% when participants viewed the relevant video segment, and by 27-35% when they did not. Efficiency increases by 10% for short videos and 25% for longer ones. However, participants tend to over-rely on AI outputs, resulting in accuracy drops of up to 32% when interacting with the deceiving AI. Alarmingly, self-reported confidence in answers remains stable across all three conditions. Our findings expose fundamental safety risks in AI-mediated video information retrieval.

When organizations delegate text generation tasks to AI providers via pay-for-performance contracts, expected payments rise when evaluation is noisy. As evaluation methods become more elaborate, the economic benefits of decreased noise are often overshadowed by increased evaluation costs. In this work, we introduce adaptive contracts for AI delegation, which allow detailed evaluation to be performed selectively after observing an initial coarse signal in order to conserve resources. We make three sets of contributions: First, we provide efficient algorithms for computing optimal adaptive contracts under natural assumptions or when core problem dimensions are small, and prove hardness of approximation in the general unstructured case. We then formulate alternative models of randomized adaptive contracts and discuss their benefits and limitations. Finally, we empirically demonstrate the benefits of adaptivity over non-adaptive baselines using question-answering and code-generation datasets.

We review thirteen generative systems and five supporting datasets for quantum circuit and quantum code generation, identified through a structured scoping review of Hugging Face, arXiv, and provenance tracing (January-February 2026). We organize the field along two axes: artifact type (Qiskit code, OpenQASM programs, circuit graphs); crossed with training regime (supervised fine-tuning, verifier-in-the-loop RL, diffusion/graph generation, agentic optimization); and systematically apply a three-layer evaluation framework covering syntactic validity, semantic correctness, and hardware executability. The central finding is that while all reviewed systems address syntax and most address semantics to some degree, none reports end-to-end evaluation on quantum hardware (Layer 3b), leaving a significant gap between generated circuits and practical deployment. Scope note: quantum code refers throughout to quantum program artifacts (QASM, Qiskit); we do not cover generation of quantum error-correcting codes (QEC).