Ilan Mandel, Madiha Zahrah Choksi, Joey Castillo, Sebastian Benthall, James Grimmelmann
Generative AI is putting the law of reverse engineering (RE) under severe strain. Existing IP and technology laws assume that RE is expensive and time-consuming, and that it requires specialized expertise. They generally allow it, but they rely on its difficulty as a filter to separate (permissible) RE from (impermissible) copying and theft. Generative AI is changing this equation by making \RE much faster, cheaper, and easier. AI systems can identify hardware components, extract firmware, decompile object code, and clone entire apps. The resulting struggle between technology vendors and reverse engineers will challenge a wide range of laws, including trade secret, copyright, anti-circumvention, computer misuse, and contract. In this article, we describe the ways in which AI is changing \RE, map these new practices onto existing laws, and sketch ways that the legal system could respond and their associated trade-offs.
Historical maps contain dense morphological and semantic information about past urban environments, but many digitization workflows focus primarily on extracting geometry, and do not capture cartographic semantics such as building use and construction materials. This study introduces ArchiveLens, a generative spatial knowledge framework that operationalizes historical maps as entity-centric, machine-queryable geospatial databases. The pipeline decouples morphological and semantic reconstruction through three stages: tiled YOLOv11-seg instance segmentation for building footprint extraction; deterministic material classification via hue, saturation, and value (HSV)-based color thresholds; and localized attribute parsing using a quantized vision-language model constrained by domain-specific prompts. Extracted entities are integrated into a hybrid Spatial-Temporal Retrieval-Augmented Generation (ST-RAG) architecture that combines dense semantic retrieval with symbolic spatio-temporal filtering and mandatory provenance grounding. Applied to 445 georeferenced Sanborn map sheets of Chicago (1918-1950), the framework achieves high geometric and semantic accuracy. Beyond performance, ArchiveLens advances GIScience by formalizing historical cartography as structured spatial knowledge representation, allowing researchers to query building attributes and compare spatial patterns across map sheets and years.
Ignatius Ezeani, Adrian Friday, Abdessalam Elhabbash, John Vidler, Daniel King, Paul Dempster
This volume contains the proceedings of the 2nd International Workshop on Low Carbon Computing (LOCO 2026), held at Lancaster University, United Kingdom, on 10-11 September 2026. LOCO provides an interdisciplinary forum for research, practical tools, early-stage work, radical ideas, and critical perspectives addressing the reduction of greenhouse gas emissions associated with computing. The proceedings cover topics including carbon measurement and reporting, sustainable software engineering, energy-efficient computing, carbon-aware systems, hardware longevity, embodied carbon, circular computing, resource management, frugal and sufficiency-oriented computing, sustainable artificial intelligence, scientific computing, and the wider environmental effects of digital technologies. Full workshop papers were evaluated through a non-blind peer-review process by members of the LOCO 2026 Programme Committee. Submissions were assessed for originality, relevance to the workshop, quality, and potential to stimulate constructive discussion. The accepted lightning-talk contributions are published separately as a companion volume, LOCO 2026 Lightning Talk Abstracts: 2nd International Workshop on Low Carbon Computing (arXiv:2609.13355).
Ignatius Ezeani, Adrian Friday, Abdessalam Elhabbash, John Vidler, Daniel King, Paul Dempster
This volume contains the accepted lightning-talk contributions from the 2nd International Workshop on Low Carbon Computing (LOCO 2026), held at Lancaster University, United Kingdom, on 10-11 September 2026. The collection brings together 14 short papers presenting emerging research, early-stage results, practical tools, and critical perspectives on reducing the environmental impacts of computing. Topics include carbon and energy measurement, carbon-aware scheduling, sustainable artificial intelligence and machine learning, high-performance computing, computational sufficiency, hardware longevity, interoperability, frugal computing, and post-growth approaches to digital technology. The contributions were selected by the LOCO 2026 Programme Committee for their relevance to the workshop and their potential to stimulate discussion across technical, organisational, and societal dimensions of low-carbon computing. They are published together in this edited volume as a companion collection to the main LOCO 2026 proceedings index.
Leandros Maglaras, George Boustras, Azidine Guezzaz, Naghmeh Moradpoor
We evaluate ErouVe, a vehicular eco-routing protocol in which roadside units (RSUs) collect per-segment travel-time and CO2 reports from passing vehicles and answer route requests with the lowest-cost alternative, across a range of road architectures and traffic scenarios. We conduct the evaluation with a working implementation of the protocol inside Veins, adapted to drive SUMO's own router rather than a hand-rolled shortest-path search, on a network offering a short, signal-controlled route and a longer, signal-free bypass, under two traffic scenarios. This evaluation reveals clear strengths and weaknesses of the protocol. Under light, uncongested traffic, ErouVe's fleet-wide effect is light. Under traffic dense enough to stress the signalized route's throughput and produce a real, growing queue, the same protocol succeeds a medium fleet-wide CO2 reduction, with diverted vehicles individually saving even more fuels. We use these findings to propose three targeted corrective methods; a knowledge-gated cost-acceptance rule, a diversion-locking rule, and a distance-normalized CO2 cost function, that respectively address a cold-start knowledge asymmetry, decision flapping under noisy crowd-sourced data, and a per-second CO2 metric that rewards idling. Each of these weaknesses are generic pitfalls for any crowd-sourced routing signal and all proposed methods measurably restore protocol effectiveness.
Jie Li, Bin Guo, Georgina Curto, Zikang Liu, Mengqi Chen, yu li
One-person companies (OPCs) are emerging enterprises operated by individual entrepreneurs through digital platforms and LLM-based AI agents. These highly skilled entrepreneurs are characterized by being digital nomads, with limited reliance on local ecosystems. Governments in China use distinct regional support policies to attract mobile OPCs, turning local support into a cross-regional coordination problem. OPCs compare regions to maximize their own utility, but imperfect policy perceptions may cause them to reject otherwise suitable locations. Existing policy-learning methods optimize resources under fixed perceptions, while consensus-formation methods aggregate views without feeding cognitive changes back into policy learning. This disconnect produces mismatched location choices and inefficient regional policy adjustment. We introduce PolicyMediator, a cognition-constrained leader-multi-follower Stackelberg framework with bilateral mediation. The mediator explains verified policies to OPCs to build awareness and trust, and reports their needs and unresolved concerns to governments for policy learning. Its self-evolving skills retain, revise, or prune strategies using verification, allocation, capacity, and objective trajectories rather than mediation success alone. Across six-region simulations in China with five OPC types, five scenarios, and three OPC-simulator backbones, PolicyMediator ranks first in all 15 OPC-backbone--scenario combinations. Over the eight-baseline mean, it improves the aggregate government objective, combining economic performance, employment benefit, and policy-access equality, by 42.4% and average OPC utility by 40.9%.
The integration of large language models (LLMs) into clinical decision-making raises unresolved questions regarding diagnostic bias. We investigate whether a documented psychiatric history alters LLM recommendations when the current clinical evidence is identical. Using a counterfactual experiment inspired by Rosenhan's 1973 field study, we factorially varied diagnostic-history condition (no prior diagnosis stated vs. prior schizophrenia diagnosis) and current-evidence condition (voices ceased vs. voices persisted) within a clinical vignette. Across five LLMs (GPT-5.6 Sol, Claude Sonnet 5, Gemini 3.6 Flash, Grok 4.6, and Mistral Large 2512), we generated 20,000 model responses across independent queries. When evidence supported discharge, adding a prior schizophrenia diagnosis increased recommendations for continued hospitalization nearly fivefold, from 4.7% to 22.3% (risk difference, 17.5 percentage points; 95% CI, 16.2-18.8). This shift occurred across all five models (risk-difference range, 10.4-34.4 percentage points). In the same condition, reference concordance decreased from 95.3% to 77.7%, while stated decision confidence increased. Across all model-by-evidence contrasts, confidence increases corresponded to Cohen's d values of 0.35-1.09. Signal-detection analyses showed that the prior diagnosis shifted the decision criterion toward hospitalization across all five models and reduced discrimination between the two current-evidence conditions in three of them. When current evidence supported continued hospitalization, the label increased the recommendation for hospitalization from 92.6% to 97.0%.
Peter Kokol, Martha Kachweka, Helena Blazun Vosner, Samuel Chol Buol, Jernej Zavrsnik, Bojan Zlahtic, Grega Zlahtic
While climate change disproportionately impacts African public health, AI predictive modeling remains heavily constrained. Using a Synthetic Knowledge Synthesis approach and VOSviewer, we conducted a bibliometric analysis of Scopus-indexed literature, comparing Global and African corpora. Network analysis reveals stark divergences: global research emphasizes complex, predictive models for planetary health, whereas African research focuses on localized, reactive solutions for immediate infectious disease triage. Severe structural barriers in Africa, including compute deficits, data silos, and energy instability, hinder advanced AI scaling. Ultimately, international funding must shift from siloed pilot projects toward building foundational digital infrastructure and localized AI frameworks.
The sustainability of the Web has drawn growing attention, as large language model (LLM)-based web services leave a substantial carbon footprint. These services typically run in geo-distributed data centers whose grid carbon intensity varies across regions and over time, so routing each query to a cleaner region or a smaller model can reduce emissions. A conversation, however, is a stateful web session: each turn reuses the key--value (KV) cache of its history, which resides only where earlier turns were served. Serving a turn on another deployment requires recomputing this history, whose carbon cost grows with the conversation and can outweigh the saving of the move. Existing carbon-aware LLM routers treat a conversation's queries as independent and ignore this cost. We present CarbonRouter, an online framework that jointly selects the model, GPU type, and region for each query to minimize carbon emissions. CarbonRouter introduces a cache-aware carbon cost model, which charges history recomputation to every candidate deployment using cached-token counts from the serving backends, and uses a lightweight gate that keeps a session on its cached deployment when it would likely move back at the next turn and rebuild its history again. We evaluate CarbonRouter in trace-driven experiments on two real-world conversation datasets with five LLMs, three GPU types, and four regions, where it achieves the lowest carbon of all policies. CarbonRouter reduces CO$_2$ by 16.1\% on average relative to FairCarbon, a state-of-the-art carbon-aware router with the same action space, while also improving predicted answer quality and estimated latency. Moreover, the saving rises further to 23.1\% for sessions longer than 30 turns.
Elena Senger, Yuri Campbell, Rob van der Goot, Barbara Plank
Large language models (LLMs) are increasingly used for codebook-based content analysis in computational social science, where codebooks translate theoretical concepts into labeling schemes. While prior work reports strong predictive performance, it remains unclear whether these systems produce reliable and robust text annotations. We conduct a large-scale empirical failure-mode analysis of LLM-based coding pipelines across four datasets and five models by systematically evaluating their stability, robustness, and conceptual validity under controlled perturbations, domain shifts, and optimization procedures. We uncover multiple failure modes that are largely invisible under standard metrics, showing that high aggregate scores frequently coexist with severe label-level disparities, instability across runs, domain sensitivity, and shifts in label interpretation. Importantly, these failure modes persist under widely used optimization strategies and remain largely invisible under standard metrics such as accuracy and micro-F1. Our results demonstrate that current LLM-based coding pipelines cannot be treated as stable measurement tools without targeted validation beyond aggregate performance.
This article reports on the third iteration of a survey of computerized tools and technologies taught as part of postgraduate translation training programmes. While the survey was carried out under the aegis of the EMT Network, more than half of responses are from outside that network. The results show the responsiveness of programmes to innovations in translation technology, with increased compulsory inclusion of machine translation, post-editing, and quality evaluation, and a rapid response to the release of generative tools. The flexibility required during the Covid-19 pandemic has also led to some lasting changes to programmes. While the range of tools being taught has continued to expand, programmes seem to be consolidating their core offering around cloud-based software with cost-free academic access. There has also been an increase in the embedding of professional contexts and workflows associated with translation technology. Generic file management and data security skills have increased in perceived importance, and legal and ethical issues related to translation data have also become more prominent. In terms of course delivery the shift away from conventional labs identified in EMT2017 has accelerated markedly, no doubt partly driven by the pandemic, accompanied by a dramatic expansion in the use of students' personal devices.
Using Large Language Models (LLMs) to give educational feedback to students for their assignments has attracted much attention in the AI in Education (AIED) field. Yet, there is currently no large-scale open-source dataset of student assignments that includes detailed assignment descriptions, rubrics, and student submissions across various courses. As a result, research on generalisable methodology for automatic generation of effective and responsible educational feedback remains limited. In this paper, we introduce a synthetic computer science university assignment dataset for LLM-based educational feedback research, called SCALEFeedback (Synthetic Computer science Assignments for LLM Educational Feedback Research). The dataset is generated via Sophisticated Assignment Mimicry (SAM) framework specifically designed to synthesise this dataset and that utilizes one-to-one LLM-based imitation from real assignment descriptions, rubrics, and student submissions. Our open-source dataset contains 10,000 synthetic student submissions spanning 155 assignments across 59 university-level computer science courses. Technical validation confirmed that the synthetic dataset closely resembles real data while successfully eliminating personally identifiable information present in the source material. The creation of this dataset is a valuable contribution to researchers who aim to develop LLM-based generalisable methods for offering high-quality, automated educational feedback in a scalable way.
Alan Chan, Christoph Winter, Andrew Barto, Jakub Pachocki, Geoffrey Hinton, Eric Horvitz, Yoshua Bengio, Dawn Song 等 22 位作者
In contrast to even a year ago, AI systems now write most of the code inside the companies that build them. As more of the AI research and development (R&D) pipeline is automated, could AI progress radically accelerate in an "intelligence explosion," where years of advances are compressed into months or less? Preliminary evidence suggests that it could. In this work, we assess this evidence, analyze an intelligence explosion's potential impacts, and propose policy responses. AI systems are on track to automate most AI R\&D work within a few years, and possibly all of it. If this triggers an intelligence explosion, it could dramatically bring forward AI's benefits, but also pose extreme risks: capabilities growth could accelerate far beyond what society can keep up with, humanity could lose control over superhuman AI systems, and checks on power within and between states, companies, and branches of government could be severely eroded. Although there remains much uncertainty about these possibilities, the high stakes warrant serious further attention. Policymakers should urgently obtain more visibility into the automation of AI R&D, develop ways to steer and constrain an intelligence explosion, and prepare society to adapt to an intelligence explosion's impacts.
Financial instruction-data pipelines must distinguish content validity from permission to release a particular dataset. Arya Forge combines typed records, current-content checks, and approval coverage bound to a dataset manifest. We evaluate its release boundaries in a financial-data systems case study. An integrity audit covers 100,000 synthetic instruction records; a separate arithmetic implementation agrees with all 15,000 numeric records across twelve formula families. In 240 scenarios with 2,400 treatment rows, adding financial and release-record checks to a short-instruction filtering recipe rejects all injected numeric faults and policy holds while retaining 220 of 240 clean controls. The twenty lost controls comprise every tested cash-conversion-cycle example. Nineteen persisted-release cases show that eleven selected negative cases are blocked, four equivalent representations are rejected, and changes to external evidence after binding are outside ordinary Gold promotion's checks. Local processing measurements and controlled checkpoint resumptions characterize cost and recovery through 100,000 input rows. The results expose a tradeoff between strict producer-contract enforcement and useful-representation retention, and identify the protected inputs required for manifest-bound approval. These findings apply to the tested implementation and workloads; they do not establish general financial correctness, authenticated approval, expert-validated quality, or downstream model improvement.
Junfeng Jiao, Kevin Chen, Shrestha Mishra, Abhejay Murali, David Atkinson, Shriti Raj, Saleh Afroogh
AI systems are increasingly present in children's everyday environments: education, entertainment, social networks, and toys. Their safety is typically assessed at the model level, through benchmarks, red-teaming, and content filtering. However, most AI harm taxonomies lack child-specific evidence and under-specify developmental harm and the stakeholder asymmetries unique to minors. Building on prior taxonomies, we construct a child-centered harm taxonomy mapped to sociotechnical failure modes. We curate and code 1331 publicly reported child-related AI incidents, annotating internal and external harm types and the failure mode implicated in each incident. Our analysis reveals which harms are most prevalent in public reporting, how individual-level and population-level harms relate, and which failures are most often coded alongside harm. We find that harm concentrates in representational, dignity-related, and psychological categories, and co-occurs most often with deployment-context, governance, and misuse failures rather than model defects. Harm to an individual child is also frequently linked to population-level harm, reaching children who never interacted with the system. These findings suggest that model-level evaluations, while necessary, cannot by themselves anticipate the harm pathways most often documented for children. We conclude with actionable implications emphasizing mitigations beyond model-level safeguards such as content filtering: age assurance, consent verification for depiction, safety-by-design UX constraints, incident disclosure norms, and auditing regimes targeting developmental risk. Privacy-minimized annotations, a codebook, and documentation are available at https://github.com/The-Responsible-AI-Initiative/KidAID.
Refusal rates conflate upstream provider behavior with downstream safeguards and do not reveal what information remains in a delivered answer. Safeguard-conditioned uplift is an evidence-gated framework for evaluating benchmark utility and risk-related proxies along deployed access paths for dual-use biology assistants. It attributes interventions to their source, selects thresholds under a legitimate-access budget, and supports content claims only when answer verifiers pass frozen qualification tests. In a prospectively frozen route and generation study, two of six configurations pass the LAB-Bench-calibrated action-level rule, but both inherit the same Claude Opus provider effect rather than independent downstream gains. Positive selectivity persists on 104 previously unused matched pairs, but benign intervention exceeds the $20\%$ limit. No joint-scoring verifier meets the frozen criteria; a factorial follow-up shows that criterion isolation and ordinal outputs improve end-to-end edit-direction accuracy on controlled template edits. The evidence supports action-level selectivity and verifier interface effects, but not calibrated access, verified content removal, or biological-risk reduction.
Peter Romero, Haotian Li, Hugo Save, Luning Sun, Zachary R. Tidler, Kazuya Horibe, Joe Watson, Daniel Romero-Alvarado 等 11 位作者
We release a battery of 28 propensities for AI research, with expert-written rubrics, synthetic datasets, and annotated demand intervals, built with the method of Romero-Alvarado et al. (2026). We selected a wide collection of generic propensities that may influence the behavior of AI systems and humans. They are inspired by psychometric, psychological, and behavioral research, so that experimental results can be understood in the context of familiar constructs previously used also for human behavior research. The rubrics are useful for annotating existing benchmarks whose items may be relevant or not for one or more propensities. For easier profiling, however, the synthetic datasets are defined in such a way that they should be sensitive to one propensity only, allowing for a quick prototyping of propensity profiles using the battery of datasets. We also outline directions for future propensity research.
Artificial intelligence (AI) is changing how data-driven research is produced. As AI tools make information and technical execution more accessible, mentorship must concentrate on judgment, ethics, and the design of end-to-end AI-augmented workflows. Disciplinary experts need mentoring to adopt AI while preserving their expertise and human agency, and AI-native research trainees need mentoring to become polymathic workflow developers. This chapter proposes Everyone's Backyard, an interdisciplinary model in which disciplinary experts, trainees, and community partners mentor one another in both directions and across disciplines through shared research practice in a convergence studio. The model centers on four capacities: statistical thinking, data ethics, collaboration, and AI for social good and society. We illustrate the model with the Spring 2026 aiX Faculty Fellows Program at Columbia University, which brought together 20 faculty fellows and 15 student interns from 14 schools and departments. Evaluation data show growth in faculty confidence, intern competence, and cross-disciplinary connection.
Search engines increasingly display AI-generated answers above organic links, potentially displacing traffic to upstream publishers. We estimate the impact of Google's AI Overviews (AIO) on Wikipedia's search traffic using AIO's staggered geographic rollout and Wikipedia's multilingual structure. Our difference-in-differences design compares monthly external-search referrals to English Wikipedia articles with referrals to the same articles in German and French, and finds that default AIO availability reduced English search traffic by 5.45% and 4.82%, respectively. Our results suggest that answer-producing digital intermediaries can materially reallocate attention away from informational publishers, with implications for content monetization, search platform design, and policy.
In September 2026, a new large language model was released. Soon after, a new way of making explainer videos became very popular: people asked the model to write a program, and the program produced the video. "One-click video" content then appeared in large amounts on many platforms. On the surface, this is a rise in creative power at the level of the individual. We argue that, at a deeper level, it changes the structure of the content industry. This judgment rests on two observations. First, this wave of content is service-type content. People watch it to meet a need, not to enjoy the content itself. The other kind is work-type content. AI makes both kinds faster to produce. But anyone can come in and make service-type content, while work-type content is limited by the number of people who have something to say. So the ratio between the two slowly becomes unbalanced. Second, free platforms choose content through layers of trial viewing, and this uses up user attention. The total amount of attention is fixed. The unbalanced ratio also dilutes good content. A simplified mathematical model shows that, as long as these conditions hold, the recommendation pool mechanism will fail once new content grows past a certain multiple. Based on this, we discuss one possible way out. Machines and entry rules judge whether service-type content passes the bar. Trusted people judge work-type content. We call this the BoLe system. At the end, we briefly imagine how the two kinds of content may develop from here.
Planetary Geospatial Foundation Models: A New Paradigm for Global Public Health
Arbaaz Muslim, Aviv Slobodkin, Katherine Wheeler-Martin, Eric Zhou, John Brittain, Martin Frasch, Hamsa Subramaniam, Jacob Bien 等 38 位作者
AI 速览：提出并展示地理空间基础模型 PDFM 如何以可泛化的地点表征补充传统流行病学工作流，并在多国多任务案例中改善公共卫生监测。
The efficacy of traditional disease prediction is limited by spatial gaps and temporal lags, which impact the timing and targets of resource deployments. Outbreaks escalate undetected, chronic disease burdens are quantified years later, and at-risk populations in data-sparse regions remain unaddressed. Planetary geospatial foundation models complement existing epidemiological workflows to provide operational improvements, encoding multimodal search, mobility, and environmental signals into generalizable place representations. As illustrations of this complementarity, we present independent global health case studies of Google Earth AI's Population Dynamics Foundation Model (PDFM) -- a foundation model for geospatial inference -- across four domains (vaccine-preventable, communicable, noncommunicable, maternal mental health), five tasks (spatial extrapolation, interpolation/nowcasting, probabilistic forecasting, prospective forecasting, risk stratification), and four countries (USA, Canada, Mexico, and the Democratic Republic of the Congo). Across these case studies, PDFM addresses critical surveillance gaps across domains: improving US-Canada border MMR vaccination coverage predictions by capturing cross-border behavioral spillovers domestic models miss; nowcasting cardiovascular disease to accelerate data availability; enhancing short-term municipal Mexican dengue forecasts for timely outbreak vector control; improving forecasts of cholera hotspots; and adding a transferable signal to individual-level postpartum-depression risk prediction in US states the model had never seen, while not replacing individual socioeconomic data or closing demographic screening gaps. Together, these results showcase capabilities of geospatial foundation models for public health surveillance.
Computer-use agents (CUAs) are commonly evaluated against fixed task specifications, but users may omit requirements, add constraints, or require approval before consequential actions. Completing the benchmark instruction does not guarantee serving the person who asked. We introduce PersonaCUABench, which holds 74 tasks across 11 domains fixed and varies only who asks. A persona agent, playing one of 44 personas in four groups, rewrites each instruction in that user's own words and answers the CUA's questions, while the CUA sees only the dialogue, never the profile or the original instruction. Across nine backbones (3,330 runs), mean task completion falls by 36% relative to the original requests. The losses have two sources: in 59.9% of cells that lose points, the agent engages with what the user asked for and still gets it wrong, and 11.2% are stops the user requested that the checker scores as failures. Even when a persona request is fully completed, the user gives the top satisfaction rating in only 62.7% of cells, against 84.1% for the original requests. CUAs should therefore be evaluated not only on task completion but also on whether they meet the requirements and constraints users express. Code: https://github.com/MatrAIx-ai/PersonaCUABench. Project page: https://matraix.ai/research/persona-cua.html
AI 速览：提出“learning interactives”框架，借助Gemini的生成式UI能力自动构建定制化互动模拟，并通过教育者初步研究与专家教师的教学评分进行评估。
It is well established that students learn best through active engagement, particularly via interactive experiences tied to real-world concepts. Creating such experiences, however, typically requires labor-intensive design, costly development, and continuous teacher guidance to achieve measurably effective outcomes. Recent advances in generative UI open the door for a future of practice where the creation of tailored interactive experiences is automated. However, off-the-shelf models lack optimization for pedagogical precision and learning principles, often yielding unsatisfactory results for complex pedagogical constructs. To address these limitations, we present learning interactives, a structured framework leveraging the generative UI capabilities of Gemini to construct effective customized interactive simulations. We evaluate our approach through both an initial study with educators as well as a pedagogical rating of a varied set of simulations by expert teachers.
Large language models (LLMs) and medical AI tools are routinely used by physicians and patients for medical advice, yet their clinical safety profiles remain poorly characterized. We present NOHARM (Numerous Options Harm Assessment for Risk in Medicine), a 1,100-task benchmark of primary care-to-specialist consultation cases to measure the frequency and severity of potentially harmful errors from LLM-generated medical consultation recommendations. NOHARM covers 10 specialties, with 12,747 expert annotations for 4,249 clinical management options. Across 20 notable LLMs and 4 widely used retrieval-augmented generation (RAG) clinical AI tools, direct application of recommendations carried potential for severe harm in up to 24.6% of cases, with errors of omission accounting for more than 80% of severe errors. Harm potential was not uniform across systems, with clinical AI tools outperforming generalist LLMs, and multi-agent AI teaming further improving performance in generalist models. In a randomized study of 101 U.S.-licensed generalist physicians, AI assistance improved physician performance compared to conventional resources. However, AI-assisted physicians frequently omitted valuable AI-generated recommendations and still scored lower than many AI systems alone. Had those recommendations been incorporated, combined human-AI responses would have outperformed both the human and AI system as used, suggesting complementary strengths and unrealized potential in human-AI teaming. Collectively, these results show that despite strong performance on medical knowledge benchmarks, widely used AI tools can produce medical consultation advice with the potential for severe harm, and highlight the need for explicit measurement of clinical safety. The benchmark and leaderboard are publicly available to support ongoing evaluation and improvement of AI systems used for clinical care.
Towards Automating Scientific Review with Google's Paper Assistant Tool
Rajesh Jayaram, Drew Tyler, David Woodruff, Corinna Cortes, Yossi Matias, Vahab Mirrokni, Vincent Cohen-Addad
AI 速览：提出AI与人类在科学评审中协作的四级分类框架，并介绍用于深度科学评审与验证的智能体框架Paper Assistant Tool（PAT）。
Artificial intelligence is driving a revolution in scientific discovery, accelerating everything from hypothesis generation to mathematical theorem proving. However, this rapid acceleration is creating a systemic challenge: traditional human peer review cannot scale to match the influx of AI-assisted science. Ultimately, to resolve this tension, we must also deploy AI to accelerate the verification and review process itself. To frame the discussion around this transition, we propose a taxonomy consisting of four progressive levels of AI-human collaboration in scientific evaluation, and discuss various trade-offs involved with each. As a step toward this future, we introduce the Paper Assistant Tool (PAT), an agentic AI framework built for deep scientific review and verification. PAT ingests full scientific manuscripts and produces a comprehensive evaluation, checking theoretical results, validating experiments, suggesting improvements, and identifying potential flaws. By utilizing inference scaling techniques, PAT is able to identify deeper issues than a single model call alone, achieving a 34% improvement over zero-shot recall on mathematical errors in the SPOT benchmark. Pilot deployments of PAT as a pre-submission tool for authors at two major Computer Science conferences -- STOC and ICML -- demonstrate its ability to identify critical errors and suggest substantive improvements to research papers. By catching errors early, PAT eases the cognitive burden placed on referees, while preserving their control over the outcomes of the review process.
Canfer Akbulut, Rasmi Elasmar, Abhishek Roy, Anthony Payne, Priyanka Suresh, Lujain Ibrahim, Seliem El-Sayed, Charvi Rastogi 等 12 位作者
Interest in the concept of AI-driven harmful manipulation is growing, yet current approaches to evaluating it are limited. This paper introduces a framework for evaluating harmful AI manipulation via context-specific human-AI interaction studies. We illustrate the utility of this framework by assessing an AI model with 10,101 participants spanning interactions in three AI use domains (public policy, finance, and health) and three locales (US, UK, and India). Overall, we find that that the tested model can produce manipulative behaviours when prompted to do so and, in experimental settings, is able to induce belief and behaviour changes in study participants. We further find that context matters: AI manipulation differs between domains, suggesting that it needs to be evaluated in the high-stakes context(s) in which an AI system is likely to be used. We also identify significant differences across our tested geographies, suggesting that AI manipulation results from one geographic region may not generalise to others. Finally, we find that the frequency of manipulative behaviours (propensity) of an AI model is not consistently predictive of the likelihood of manipulative success (efficacy), underscoring the importance of studying these dimensions separately. To facilitate adoption of our evaluation framework, we detail our testing protocols and make relevant materials publicly available. We conclude by discussing open challenges in evaluating harmful manipulation by AI models.
AI 速览：提出人口动态基础模型（PDFM），通过图神经网络融合多源地理数据生成嵌入，在27个地理空间任务上取得领先表现。
Supporting the health and well-being of dynamic populations around the world requires governmental agencies, organizations and researchers to understand and reason over complex relationships between human behavior and local contexts in order to identify high-risk groups and strategically allocate limited resources. Traditional approaches to these classes of problems often entail developing manually curated, task-specific features and models to represent human behavior and the natural and built environment, which can be challenging to adapt to new, or even, related tasks. To address this, we introduce a Population Dynamics Foundation Model (PDFM) that aims to capture the relationships between diverse data modalities and is applicable to a broad range of geospatial tasks. We first construct a geo-indexed dataset for postal codes and counties across the United States, capturing rich aggregated information on human behavior from maps, busyness, and aggregated search trends, and environmental factors such as weather and air quality. We then model this data and the complex relationships between locations using a graph neural network, producing embeddings that can be adapted to a wide range of downstream tasks using relatively simple models. We evaluate the effectiveness of our approach by benchmarking it on 27 downstream tasks spanning three distinct domains: health indicators, socioeconomic factors, and environmental measurements. The approach achieves state-of-the-art performance on all 27 geospatial interpolation tasks, and on 25 out of the 27 extrapolation and super-resolution tasks. We combined the PDFM with a state-of-the-art forecasting foundation model, TimesFM, to predict unemployment and poverty, achieving performance that surpasses fully supervised forecasting. The full set of embeddings and sample code are publicly available for researchers.
Tony Feng, Trieu H. Trinh, Garrett Bingham, Dawsen Hwang, Yuri Chervonyi, Junehyuk Jung, Joonkyung Lee, Carlo Pagano 等 28 位作者
Recent advances in foundational models have yielded reasoning systems capable of achieving a gold-medal standard at the International Mathematical Olympiad. The transition from competition-level problem-solving to professional research, however, requires navigating vast literature and constructing long-horizon proofs. In this work, we introduce Aletheia, a math research agent that iteratively generates, verifies, and revises solutions end-to-end in natural language. Specifically, Aletheia is powered by an advanced version of Gemini Deep Think for challenging reasoning problems, a novel inference-time scaling law that extends beyond Olympiad-level problems, and intensive tool use to navigate the complexities of mathematical research. We demonstrate the capability of Aletheia from Olympiad problems to PhD-level exercises and most notably, through several distinct milestones in AI-assisted mathematics research: (a) a research paper (Feng26) generated by AI without any human intervention in calculating certain structure constants in arithmetic geometry called eigenweights; (b) a research paper (LeeSeo26) demonstrating human-AI collaboration in proving bounds on systems of interacting particles called independent sets; and (c) an extensive semi-autonomous evaluation (Feng et al., 2026a) of 700 open problems on Bloom's Erdos Conjectures database, including autonomous solutions to four open questions. In order to help the public better understand the developments pertaining to AI and mathematics, we suggest quantifying standard levels of autonomy and novelty of AI-assisted results, as well as propose a novel concept of human-AI interaction cards for transparency. We conclude with reflections on human-AI collaboration in mathematics and share all prompts as well as model outputs at https://github.com/google-deepmind/superhuman/tree/main/aletheia.
AI 速览：随机实验发现AI辅助会损害开发者对新编程库的概念理解、代码阅读和调试能力，且平均未带来显著效率提升，但某些认知参与型交互模式可保留学习效果。
AI assistance produces significant productivity gains across professional domains, particularly for novice workers. Yet how this assistance affects the development of skills required to effectively supervise AI remains unclear. Novice workers who rely heavily on AI to complete unfamiliar tasks may compromise their own skill acquisition in the process. We conduct randomized experiments to study how developers gained mastery of a new asynchronous programming library with and without the assistance of AI. We find that AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average. Participants who fully delegated coding tasks showed some productivity improvements, but at the cost of learning the library. We identify six distinct AI interaction patterns, three of which involve cognitive engagement and preserve learning outcomes even when participants receive AI assistance. Our findings suggest that AI-enhanced productivity is not a shortcut to competence and AI assistance should be carefully adopted into workflows to preserve skill formation -- particularly in safety-critical domains.
LearnLM Team, Google, :, Alicia Martín, Amir Globerson, Amy Wang, Anirudh Shekhawat, Anna Iurchenko 等 37 位作者
Textbooks are a cornerstone of education, but they have a fundamental limitation: they are a one-size-fits-all medium. Any new material or alternative representation requires arduous human effort, so that textbooks cannot be adapted in a scalable manner. We present an approach for transforming and augmenting textbooks using generative AI, adding layers of multiple representations and personalization while maintaining content integrity and quality. We refer to the system built with this approach as Learn Your Way. We report pedagogical evaluations of the different transformations and augmentations, and present the results of a a randomized control trial, highlighting the advantages of learning with Learn Your Way over regular textbook usage.