In 2023, the National Eating Disorders Association's (NEDA) chatbot Tessa was suspended after providing harmful weight-loss advice to vulnerable users--an avoidable failure that underscores the risks of unsafe AI in healthcare contexts. This paper examines Tessa as a case study in absent safety engineering and demonstrates how a lightweight, modular safeguard could have prevented the incident. We propose a hybrid safety middleware that combines deterministic lexical gates with an in-line large language model (LLM) policy filter, enforcing fail-closed verdicts and escalation pathways within a single model call. Using synthetic evaluations, we show that this design achieves perfect interception of unsafe prompts at baseline cost and latency, outperforming traditional multistage pipelines. Beyond technical remedies, we map Tessa's failure patterns to established frameworks (OW ASP LLM Top10; NIST SP 800-53), connecting practical safeguards to actionable governance controls. The results highlight that robust, auditable safety in health-adjacent AI does not require heavyweight infrastructure: explicit, testable checks at the last mile are sufficient to prevent "another Tessa," while governance and escalation ensure sustainability in real-world deployment.

Large Language Models (LLMs) are increasingly capable of interpreting multimodal inputs to generate complex 3D shapes, yet robust methods to evaluate geometric and structural fidelity remain underdeveloped. This paper introduces a human-in-the-loop framework for the quantitative evaluation of LLM-generated 3D models, supporting applications such as democratization of CAD design, reverse engineering of legacy designs, and rapid prototyping. We propose a comprehensive suite of similarity and complexity metrics--including volumetric accuracy, surface alignment, dimensional fidelity, and topological intricacy--to benchmark generated models against ground-truth CAD references. Using an L-bracket component as a case study, we systematically compare LLM performance across four input modalities: 2D orthographic views, isometric sketches, geometric structure trees, and code-based correction prompts. Our findings demonstrate improved generation fidelity with increased semantic richness, with code-level prompts achieving perfect reconstruction across all metrics. A key contribution of this work is demonstrating that our proposed quantitative evaluation approach enables significantly faster convergence toward the ground truth, especially compared to traditional qualitative methods based solely on visual inspection and human intuition. This work not only advances the understanding of AI-assisted shape synthesis but also provides a scalable methodology to validate and refine generative models for diverse CAD applications.

This paper introduces ArGen (Auto-Regulation of Generative AI systems), a framework for aligning Large Language Models (LLMs) with complex sets of configurable, machine-readable rules spanning ethical principles, operational safety protocols, and regulatory compliance standards. Moving beyond just preference-based alignment, ArGen is designed to ensure LLMs adhere to these multifaceted policies through a novel synthesis of principle-based automated reward scoring, Group Relative Policy Optimisation (GRPO), and an Open Policy Agent (OPA) inspired governance layer. This approach provides the technical foundation for achieving and demonstrating compliance with diverse and nuanced governance requirements. To showcase the framework's capability to operationalize a deeply nuanced and culturally-specific value system, we present an in-depth case study: the development of a medical AI assistant guided by principles from Dharmic ethics (such as Ahimsa and Dharma), as derived from texts like the Bhagavad Gita. This challenging application demonstrates ArGen's adaptability, achieving a 70.9% improvement in domain-scope adherence over the baseline. Through our open-source repository, we show that ArGen's methodology offers a path to 'Governable Al' systems that are technically proficient, ethically robust, and verifiably compliant for safe deployment in diverse global contexts.

Large Language Models (LLMs) have scaled rapidly in size and complexity, requiring increasingly intricate parallelism for distributed training, such as 3D parallelism. This sophistication motivates a shift toward simpler, more debuggable programming paradigm like Single Program Multiple Data (SPMD). However, SPMD in eager execution introduces two key challenges: ensuring consistency with single-device execution and achieving high performance at scale. In this paper, we introduce veScale, an eager-mode training system that fully embraces SPMD paradigm to democratize distributed tensor programming. veScale addresses the prevalent issue of inconsistent results in systems like PyTorch by introducing a novel algorithm of distributed Random Number Generation (RNG) compatible with arbitrary sharded operators. veScale also significantly boosts training performance by reducing PyTorch primitive's overhead and improving communication efficiency. Evaluations show that veScale delivers up to 2.2x speedup over the state-of-the-art training systems, like TorchTitan, and cuts code complexity by 78.4%, while preserving single-device-equivalent results.

--Open-source Vision-Language Models show immense promise for enterprise applications, yet a critical disconnect exists between academic evaluation and enterprise deployment requirements. Current benchmarks rely heavily on multiple-choice questions and synthetic data, failing to capture the complexity of real-world business applications like social media content analysis. This paper introduces VLM-in-the-Wild (ViLD), a comprehensive framework to bridge this gap by evaluating VLMs on operational enterprise requirements. We define ten business-critical tasks: logo detection, OCR, object detection, human presence and demographic analysis, human activity and appearance analysis, scene detection, camera perspective and media quality assessment, dominant colors, comprehensive description, and NSFW detection. T o this framework, we bring an innovative BlockWeaver Algorithm that solves the challenging problem of comparing unordered, variably-grouped OCR outputs from VLMs without relying on embeddings or LLMs, achieving remarkable speed and reliability. Besides, ViLD's methodology avoids traditional bounding boxes, which are ill-suited for generative VLMs, in favour of a novel spatial-temporal grid system that captures localisation information effectively for both images and videos. T o demonstrate efficacy of ViLD, we constructed a new benchmark dataset of 7,500 diverse samples, carefully stratified from a corpus of one million real-world images and videos. ViLD provides actionable insights by combining semantic matching (both embedding-based and LLMas-a-judge approaches), traditional metrics, and novel methods to measure the completeness and faithfulness of descriptive outputs. By benchmarking leading open-source VLMs (Qwen, MIMO, and InternVL) against a powerful proprietary baseline as per ViLD framework, we provide one of the first industry-grounded, task-driven assessment of VLMs capabilities, offering actionable insights for their deployment in enterprise environments. Vision-Language Models (VLMs) have fundamentally transformed the landscape of artificial intelligence, enabling systems to understand and reason about visual content through natural language.

As large language models (LLMs) are increasingly deployed in real-world applications, ensuring the safety of their outputs during decoding has become a critical challenge. However, existing decoding-time interventions, such as Contrastive Decoding, often force a severe trade-off between safety and response quality. In this work, we propose CARE, a novel framework for decoding-time safety alignment that integrates three key components: (1) a guard model for real-time safety monitoring, enabling detection of potentially unsafe content; (2) a rollback mechanism with a token buffer to correct unsafe outputs efficiently at an earlier stage without disrupting the user experience; and (3) a novel introspection-based intervention strategy, where the model generates self-reflective critiques of its previous outputs and incorporates these reflections into the context to guide subsequent decoding steps. The framework achieves a superior safety-quality trade-off by using its guard model for precise interventions, its rollback mechanism for timely corrections, and our novel introspection method for effective self-correction. Experimental results demonstrate that our framework achieves a superior balance of safety, quality, and efficiency, attaining a low harmful response rate and minimal disruption to the user experience while maintaining high response quality.

Large language models excel at basic reasoning but struggle with tasks that require interaction with external tools. We present RLFactory, a plug-and-play reinforcement learning post-training framework for multi-round tool use. RLFactory tackles (i) tool-call stability and adaptability amid tool heterogeneity and interface issues via an asyncio-based asynchronous caller and a decoupled tool/training architecture, and (ii) diverse evaluation needs via a reward layer supporting rule-based, model-judgment, and tool-verification signals. It reconstructs the MDP by introducing observation markers from tool feedback, closing the loop among model, tools, and environment, and implements a generate-parse-invoke-update workflow for dynamic policy optimization. On Search-R1 with Qwen3-4B, RLFactory achieves a 0.486 test score on the Natural Questions (NQ) dataset, surpassing larger models trained with similar techniques (e.g., Qwen2.5-7B-Instruct-GRPO at 0.473), and increases training throughput by 6.8x. RLFactory provides a low-barrier, highly adaptable framework for strengthening multi-round tool use of LLMs in real-world scenarios. Code: https://github.com/Simple-Efficient/RL-Factory.

High-quality, medically validated labels exist for clinical actigraphy data but not for ubiquitous consumer wearables like the Apple Watch. Manually labeling wearables data is expensive and doesn't scale. This paper offers a novel framework that transfers valuable labels from a source domain (e.g., actigraphy) to a target domain (e.g., Apple Watch) without requiring paired data. Instead of working with raw time-series signals, we project both domains into a shared latent embedding space using time-series foundation models (TSFMs) and develop a new framework to align the cross-device representations. Our method, Adversarial Alignment of TSFM Embeddings forces the distributions of source and target embeddings to align within this space, facilitating label transfer across device type.

Executing language-conditioned tasks in dynamic visual environments remains a central challenge in embodied AI. Existing Vision-Language-Action (VLA) models predominantly adopt reactive state-to-action mappings, often leading to shortsighted behaviors and poor robustness in dynamic scenes. This training scheme enhances modular reasoning and equips the model with transferable visual foresight, which is critical for complex and dynamic environments. Vision-Language-Action (VLA) models (Kim et al., 2024; Team et al., 2025a; Black et al., 2024) aim to equip robots with the ability to execute natural language instructions in visually rich environments. By aligning language instructions with perceptual inputs and mapping them to actions, such models enable language-guide manipulation and versatile human-robot interaction. However, reliable performance in realistic settings remains elusive: environments are inherently dynamic, i.e., objects move, contexts shift, and instructions unfold over time, so robots must ground ambiguous language, handle diverse objects, and maintain long-horizon temporal coherence as scenes evolve. These conditions expose a core limitation of purely reactive state-to-action mappings: without predictive foresight about likely future states, policies become short-sighted and brittle under distribution shifts. Previous efforts on manipulation policy learning can be broadly grouped into three paradigms, as illustrated in Figure 1. The earliest line of work employs only an action expert trained end-to-end from observations to low-level actions (Zhao et al., 2023; Chi et al., 2023), but such purely reactive mappings lack semantic grounding and generalization across tasks and embodiments (Figure 1(a)). The earliest end-to-end manipulation policies are illustrated in Figure 1(a), such as ACT (Zhao et al., 2023) and DP (Chi et al., 2023). There are also approaches, as seen in Figure 1(c), e.g., VPP (Hu et al., 2024) and Genie Envisioner (Liao et al., 2025b), that leverage video diffusion models to guide action execution through video prediction. As depicted in Figure 1(d), we adopts an integrated architecture of understanding, generation, and execution, empowering the action execution module with capabilities in both scene and instruction comprehension as well as dynamic temporal prediction.

Things to Do in L.A. Tap to enable a layout that focuses on the article. Voice comes from the use of AI. Please report any issues or inconsistencies here . OpenAI will roll out parental controls within the month, allowing parents to link accounts and receive alerts when the system detects "acute distress." The changes follow a California family's lawsuit after their 16-year-old son died by suicide following intimate conversations with ChatGPT about his mental health struggles.