"AI" tools are all the rage at the moment, even among users who aren't all that savvy when it comes to conventional software or security--and that's opening up all sorts of new opportunities for hackers and others who want to take advantage of them. A new research team has discovered a way to hide prompt injection attacks in uploaded images. A prompt injection attack is a way to hide instructions for an LLM or other "artificial intelligence" system, usually somewhere a human operator can't see them. It's the whispered "loser-says-what" of computer security. A great example is hiding a phishing attempt in an email in plain text that's colored the same as the background, knowing that Gemini will summarize the text even though the human recipient can't read it.

Drive down the 280 freeway in San Francisco and you might believe AI is everywhere, and everything. Nearly every billboard advertises an AI related product: "We've Automated 2,412 BDRs." "All that AI and still no ROI?" "Cheap on-demand GPU clusters." It's hard to know if you're interpreting the industry jargon correctly while zooming past in your vehicle. The signs are just one example of the tech industry's en-masse pivot to AI, a technology that the executives who have the most to gain from it say will be universe-shifting, inevitable and unavoidable. In California's tech heartland, every company is now an AI company, just like every company became a tech company sometime in the 2010s.

Elon Musk has accused Apple and OpenAI in a lawsuit of unfairly favoring the artificial intelligence company across iPhones and thwarting competition for other chatbot makers. Musk's X and xAI seek billions of dollars in damages in the suit filed Monday in U.S. federal court in Fort Worth, Texas, arguing that Apple's decision to integrate OpenAI into the iPhone's operating system inhibits rivalry and innovation within the AI industry and harms consumers by depriving them of choice. The billionaire founder of xAI, which now houses the Grok AI team and X social network, said Apple makes it impossible for anyone other than OpenAI's ChatGPT to reach the top of the App Store charts, a sought-after global spotlight for app developers.

In this PhD thesis, we propose a novel framework for uncertainty quantification in machine learning, which is based on proper scores. Uncertainty quantification is an important cornerstone for trustworthy and reliable machine learning applications in practice. Usually, approaches to uncertainty quantification are problem-specific, and solutions and insights cannot be readily transferred from one task to another. Proper scores are loss functions minimized by predicting the target distribution. Due to their very general definition, proper scores apply to regression, classification, or even generative modeling tasks. We contribute several theoretical results, that connect epistemic uncertainty, aleatoric uncertainty, and model calibration with proper scores, resulting in a general and widely applicable framework. We achieve this by introducing a general bias-variance decomposition for strictly proper scores via functional Bregman divergences. Specifically, we use the kernel score, a kernel-based proper score, for evaluating sample-based generative models in various domains, like image, audio, and natural language generation. This includes a novel approach for uncertainty estimation of large language models, which outperforms state-of-the-art baselines. Further, we generalize the calibration-sharpness decomposition beyond classification, which motivates the definition of proper calibration errors. We then introduce a novel estimator for proper calibration errors in classification, and a novel risk-based approach to compare different estimators for squared calibration errors. Last, we offer a decomposition of the kernel spherical score, another kernel-based proper score, allowing a more fine-grained and interpretable evaluation of generative image models.

Standard techniques for aligning large language models (LLMs) utilize human-produced data, which could limit the capability of any aligned LLM to human level. Label refinement and weak training have emerged as promising strategies to address this superalignment problem. In this work, we adopt probabilistic assumptions commonly used to study label refinement and analyze whether refinement can be outperformed by alternative approaches, including computationally intractable oracle methods. We show that both weak training and label refinement suffer from irreducible error, leaving a performance gap between label refinement and the oracle. These results motivate future research into developing alternative methods for weak to strong generalization that synthesize the practicality of label refinement or weak training and the optimality of the oracle procedure.

Time series foundation models (TSFMs) such as Lag-Llama, TimeGPT, Chronos, MOMENT, UniTS, and TimesFM have shown strong generalization and zero-shot capabilities for time series forecasting, anomaly detection, classification, and imputation. Despite these advantages, their predictions still suffer from variance, domain-specific bias, and limited uncertainty quantification when deployed on real operational data. This paper investigates a suite of statistical and ensemble-based enhancement techniques, including bootstrap-based bagging, regression-based stacking, prediction interval construction, statistical residual modeling, and iterative error feedback, to improve robustness and accuracy. Using the Belgium Electricity Short-Term Load Forecasting dataset as a case study, we demonstrate that the proposed hybrids consistently outperform standalone foundation models across multiple horizons. Regression-based ensembles achieve the lowest mean squared error; bootstrap aggregation markedly reduces long-context errors; residual modeling corrects systematic bias; and the resulting prediction intervals achieve near nominal coverage with widths shrinking as context length increases. The results indicate that integrating statistical reasoning with modern foundation models yields measurable gains in accuracy, reliability, and interpretability for real-world time series applications.

Large audio-language models (LALMs) extend the large language models with multimodal understanding in speech, audio, etc. While their performances on speech and audio-processing tasks are extensively studied, their reasoning abilities remain underexplored. Particularly, their multi-hop reasoning, the ability to recall and integrate multiple facts, lacks systematic evaluation. Existing benchmarks focus on general speech and audio-processing tasks, conversational abilities, and fairness but overlook this aspect. To bridge this gap, we introduce SAKURA, a benchmark assessing LALMs' multi-hop reasoning based on speech and audio information. Results show that LALMs struggle to integrate speech/audio representations for multi-hop reasoning, even when they extract the relevant information correctly, highlighting a fundamental challenge in multimodal reasoning. Our findings expose a critical limitation in LALMs, offering insights and resources for future research.

Recently, AI-driven interactions with computing devices have advanced from basic prototype tools to sophisticated, LLM-based systems that emulate human-like operations in graphical user interfaces. We are now witnessing the emergence of \emph{Computer-Using Agents} (CUAs), capable of autonomously performing tasks such as navigating desktop applications, web pages, and mobile apps. However, as these agents grow in capability, they also introduce novel safety and security risks. Vulnerabilities in LLM-driven reasoning, with the added complexity of integrating multiple software components and multimodal inputs, further complicate the security landscape. In this paper, we present a systematization of knowledge on the safety and security threats of CUAs. We conduct a comprehensive literature review and distill our findings along four research objectives: \textit{\textbf{(i)}} define the CUA that suits safety analysis; \textit{\textbf{(ii)} } categorize current safety threats among CUAs; \textit{\textbf{(iii)}} propose a comprehensive taxonomy of existing defensive strategies; \textit{\textbf{(iv)}} summarize prevailing benchmarks, datasets, and evaluation metrics used to assess the safety and performance of CUAs. Building on these insights, our work provides future researchers with a structured foundation for exploring unexplored vulnerabilities and offers practitioners actionable guidance in designing and deploying secure Computer-Using Agents.

Unlimited, or so-called helpful-only language models are trained without safety alignment constraints and never refuse user queries. They are widely used by leading AI companies as internal tools for red teaming and alignment evaluation. For example, if a safety-aligned model produces harmful outputs similar to an unlimited model, this indicates alignment failures that require further attention. Despite their essential role in assessing alignment, such models are not available to the research community. We introduce Jinx, a helpful-only variant of popular open-weight LLMs. Jinx responds to all queries without refusals or safety filtering, while preserving the base model's capabilities in reasoning and instruction following. It provides researchers with an accessible tool for probing alignment failures, evaluating safety boundaries, and systematically studying failure modes in language model safety.

LoRA-based large model parameter-efficient fine-tuning (PEFT) methods use low-rank de- composition to approximate updates to model parameters. However, compared to full- parameter fine-tuning, low-rank updates often lead to a performance gap in downstream tasks. To address this, we introduce DropLoRA, a novel pruning-based approach that focuses on pruning the rank dimension. Unlike conven- tional methods that attempt to overcome the low-rank bottleneck, DropLoRA innovatively integrates a pruning module between the two low-rank matrices in LoRA to simulate dy- namic subspace learning. This dynamic low- rank subspace learning allows DropLoRA to overcome the limitations of traditional LoRA, which operates within a static subspace. By continuously adapting the learning subspace, DropLoRA significantly boosts performance without incurring additional training or infer- ence costs. Our experimental results demon- strate that DropLoRA consistently outperforms LoRA in fine-tuning the LLaMA series across a wide range of large language model gener- ation tasks, including commonsense reason- ing, mathematical reasoning, code generation, and instruction-following. Our code is avail- able at https://github.com/TayeeChang/DropLoRA.