The growth and permeation of artificial intelligence (AI) technologies across society has drawn focus to the ways in which the responsible use of these technologies can be facilitated through AI governance. Increasingly, large companies and governments alike have begun to articulate and, in some cases, enforce governance preferences through AI policy. Yet existing literature documents an unwieldy heterogeneity in ethical principles for AI governance, while our own prior research finds that discussions of the implications of AI policy are not yet present in the computer science (CS) curriculum. In this context, overlapping jurisdictions and even contradictory policy preferences across private companies, local, national, and multinational governments create a complex landscape for AI policy which, we argue, will require AI developers able adapt to an evolving regulatory environment. Preparing computing students for the new challenges of an AI-dominated technology industry is therefore a key priority for the CS curriculum. In this discussion paper, we seek to articulate a framework for integrating discussions on the nascent AI policy landscape into computer science courses. We begin by summarizing recent AI policy efforts in the United States and European Union. Subsequently, we propose guiding questions to frame class discussions around AI policy in technical and non-technical (e.g., ethics) CS courses. Throughout, we emphasize the connection between normative policy demands and still-open technical challenges relating to their implementation and enforcement through code and governance structures. This paper therefore represents a valuable contribution towards bridging research and discussions across the areas of AI policy and CS education, underlining the need to prepare AI engineers to interact with and adapt to societal policy preferences.

We study empirical scaling laws for language model merging measured by cross-entropy. Despite its wide practical use, merging lacks a quantitative rule that predicts returns as we add experts or scale the model size. We identify a compact power law that links model size and expert number: the size-dependent floor decreases with model capacity, while the merging tail exhibits clear diminishing returns in the number of experts. The law holds in-domain and cross-domain, tightly fits measured curves across diverse architectures and methods (Average, TA, TIES, DARE), and explains two robust regularities: most gains arrive early, and variability shrinks as more experts are included. Building on this, we present a simple theory that explains why gains fall roughly as 1/k and links the floor and tail to properties of the base model and the diversity across domains. This law enables predictive planning: estimate how many experts are needed to reach a target loss, decide when to stop adding experts, and trade off scaling the base model versus adding experts under a fixed budget--turning merging from heuristic practice into a computationally efficient, planable alternative to multitask training. This suggests a scaling principle for distributed generative AI: predictable gains can be achieved by composing specialists, offering a complementary path toward AGI-level systems.

Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods--such as guardrails and tool calling--often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS (TAxonomy of COmprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS is a taxonomy that can cover a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal useful insights about train data distribution and pretrained knowledge of base models.

Reliable evaluation of large language models is essential to ensure their applicability in practical scenarios. Traditional benchmark-based evaluation methods often rely on fixed reference answers, limiting their ability to capture important qualitative aspects of generated responses. To address these shortcomings, we propose an integrated evaluation framework called \textit{self-refining descriptive evaluation with expert-driven diagnostics}, SPEED, which utilizes specialized functional experts to perform comprehensive, descriptive analyses of model outputs. Unlike conventional approaches, SPEED actively incorporates expert feedback across multiple dimensions, including hallucination detection, toxicity assessment, and lexical-contextual appropriateness. Experimental results demonstrate that SPEED achieves robust and consistent evaluation performance across diverse domains and datasets. Additionally, by employing relatively compact expert models, SPEED demonstrates superior resource efficiency compared to larger-scale evaluators. These findings illustrate that SPEED significantly enhances fairness and interpretability in LLM evaluations, offering a promising alternative to existing evaluation methodologies.

A central challenge in language models (LMs) is faithfulness hallucination: the generation of information unsubstantiated by input context. To study this problem, we propose Precise Information Control (PIC), a new task formulation that requires models to generate long-form outputs grounded in a provided set of short self-contained statements, without adding any unsupported ones. PIC includes a full setting that tests a model's ability to include exactly all input claims, and a partial setting that requires the model to selectively incorporate only relevant claims. We present PIC-Bench, a benchmark of eight long-form generation tasks (e.g., summarization, biography generation) adapted to the PIC setting, where LMs are supplied with well-formed, verifiable input claims. Our evaluation of a range of open and proprietary LMs on PIC-Bench reveals that, surprisingly, state-of-the-art LMs still hallucinate against user-provided input in over 70% of generations. To alleviate this lack of faithfulness, we introduce a post-training framework that uses a weakly supervised preference data construction method to train an 8B PIC-LM with stronger PIC ability--improving from 69.1% to 91.0% F1 in the full PIC setting. When integrated into end-to-end factual generation pipelines, PIC-LM improves exact match recall by 17.1% on ambiguous QA with retrieval, and factual precision by 30.5% on a birthplace fact-checking task, underscoring the potential of precisely grounded generation.

Post-hoc interpretability methods typically attribute a model's behavior to its components, data, or training trajectory in isolation. This leads to explanations that lack a unified view and may miss key interactions. While combining existing methods or applying them at different training stages offers broader insights, such approaches usually lack theoretical support. In this work, we present ExPLAIND, a unified framework that integrates all these perspectives. First, we generalize recent work on gradient path kernels, which reformulate models trained by gradient descent as a kernel machine, to realistic settings like AdamW. We empirically validate that a CNN and a Transformer are accurately replicated by this reformulation. Second, we derive novel parameter- and step-wise influence scores from the kernel feature maps. Their effectiveness for parameter pruning is comparable to existing methods, demonstrating their value for model component attribution. Finally, jointly interpreting model components and data over the training process, we leverage ExPLAIND to analyze a Transformer that exhibits Grokking. Our findings support previously proposed stages of Grokking, while refining the final phase as one of alignment of input embeddings and final layers around a representation pipeline learned after the memorization phase. Overall, ExPLAIND provides a theoretically grounded, unified framework to interpret model behavior and training dynamics.

Causal inference typically assumes centralized access to individual-level data. Yet, in practice, data are often decentralized across multiple sites, making centralization infeasible due to privacy, logistical, or legal constraints. We address this problem by estimating the Average Treatment Effect (ATE) from decentralized observational data via a Federated Learning (FL) approach, allowing inference through the exchange of aggregate statistics rather than individual-level data. We propose a novel method to estimate propensity scores by computing a federated weighted average of local scores with Membership Weights (MW)--probabilities of site membership conditional on covariates--which can be flexibly estimated using parametric or non-parametric classification models. Unlike density ratio weights (DW) from the transportability and generalization literature, which either rely on strong modeling assumptions or cannot be implemented in FL, MW can be estimated using standard FL algorithms and are more robust, as they support flexible, non-parametric models--making them the preferred choice in multi-site settings with strict data-sharing constraints. The resulting propensity scores are used to construct Federated Inverse Propensity Weighting (Fed-IPW) and Augmented IPW (Fed-AIPW) estimators. Unlike meta-analysis methods, which fail when any site violates positivity, our approach leverages heterogeneity in treatment assignment across sites to improve overlap. We show that Fed-IPW and Fed-AIPW perform well under site-level heterogeneity in sample sizes, treatment mechanisms, and covariate distributions. Both theoretical analysis and experiments on simulated and real-world data highlight their advantages over meta-analysis and related methods.

Ensuring Large Language Model (LLM) safety remains challenging due to the absence of universal standards and reliable content validators, making it difficult to obtain effective training signals. We discover that aligned models already possess robust internal safety beliefs: they consistently produce high-confidence refusals to harmful requests while exhibiting high entropy when generating potentially dangerous content. This entropy gap reveals an untapped signal--models intrinsically "know" when to refuse. We introduce Safety Instincts Reinforcement Learning (SIRL), which transforms this internal confidence into a self-generated reward signal, eliminating dependence on external validators or human annotations. SIRL teaches models to trust their safety instincts by reinforcing low-entropy refusal behaviors. Evaluated on Llama and Qwen models, SIRL maintains 89%+ Defense Success Rates (DSRs) against 20+ jailbreak methods, from static prompts to adaptive attacks. Using only 15,000 unlabeled prompts, SIRL surpasses resource-intensive supervised methods while preserving performance on mathematics, coding, and conversation benchmarks. Our work demonstrates that effective alignment can emerge from within, paving the way for more autonomous and robust AI safety mechanisms that scale without extensive human oversight. The widespread deployment of large language models (LLMs) has made defending against jailbreak attacks a critical priority (Yi et al., 2024; Wei et al., 2023; Shen et al., 2025b). Unlike well-defined tasks with clear metrics, determining what constitutes "safe" behavior requires expensive human annotation, carefully crafted reward models, or predefined rules that often fail to generalize (Casper et al., 2023; Zou et al., 2023b). As sophisticated jailbreak techniques continue to evolve (Samvelyan et al., 2024; Zou et al., 2023b; Chao et al., 2025; Andriushchenko & Flammarion, 2024; Andriushchenko et al., 2025), the question remains: can models learn to enhance their own safety without relying on these external validators? Recent advances in self-alignment (Burns et al., 2023; Christiano et al., 2018) and the pursuit of su-peralignment (Leike & Sutskever, 2023) suggest that models may possess untapped internal signals for improvement. Inspired by this possibility, we investigate whether aligned LLMs harbor intrinsic safety beliefs that could guide self-improvement.

Large reasoning models (LRMs) "think" by generating structured chain-of-thought (CoT) before producing a final answer, yet they still lack the ability to reason critically about safety alignment and are easily biased when a flawed premise is injected into their thought process. We propose RECAP (Robust Safety Alignment via Counter-Aligned Prefilling), a principled reinforcement learning (RL) method for post-training that explicitly teaches models to override flawed reasoning trajectories and reroute to safe and helpful responses. RECAP trains on a mixture of synthetically generated counter-aligned CoT prefills and standard prompts, requires no additional training cost or modifications beyond vanilla reinforcement learning from human feedback (RLHF), and substantially improves safety and jailbreak robustness, reduces overrefusal, and preserves core reasoning capability -- all while maintaining inference token budget. Extensive analysis shows that RECAP-trained models engage in self-reflection more frequently and remain robust under adaptive attacks, preserving safety even after repeated attempts to override their reasoning.

Abstract--Large Language Models (LLMs) are increasingly integrated into software engineering (SE) tools for tasks that extend beyond code synthesis, including judgment under uncertainty and reasoning in ethically significant contexts. We present a fully automated framework for assessing ethical reasoning capabilities across 16 LLMs in a zero-shot setting, using 30 real-world ethically charged scenarios. Each model is prompted to identify the most applicable ethical theory to an action, assess its moral acceptability, and explain the reasoning behind their choice. Responses are compared against expert ethicists' choices using inter-model agreement metrics. Our results show that LLMs achieve an average Theory Consistency Rate (TCR) of 73.3% and Binary Agreement Rate (BAR) on moral acceptability of 86.7%, with interpretable divergences concentrated in ethically ambiguous cases. A qualitative analysis of free-text explanations reveals strong conceptual convergence across models despite surface-level lexical diversity. These findings support the potential viability of LLMs as ethical inference engines within SE pipelines, enabling scalable, auditable, and adaptive integration of user-aligned ethical reasoning. Our focus is the Ethical Interpreter component of a broader profiling pipeline: we evaluate whether current LLMs exhibit sufficient interpretive stability and theory-consistent reasoning to support automated profiling. Autonomous systems are increasingly becoming an integral part of our daily lives across diverse domains [1], [2]. These systems can operate independently without any human intervention and make decisions acting on behalf of their users [3]-[6]. Their rapid growth brings both opportunities and challenges. From a software engineering perspective, as these systems become pervasive, a key challenge is designing systems that, beyond meeting technical requirements, also account for ethical considerations [7]-[11]. Recently, various studies have focused on the ethical implications of these software-intensive systems on individuals and society [10], [12]-[15]. Software engineering ethics encompasses principles and rules that guide engineers' decisions throughout the design and development process [16]. V arious approaches have also been introduced that ensure that systems align with broad ethical values like fairness, transparency, and safety [17]-[22].