Retrieval-Augmented Generation (RAG) is a well-established and rapidly evolving field within AI that enhances the outputs of large language models by integrating relevant information retrieved from external knowledge sources. While industry adoption of RAG is now beginning, there is a significant lack of research on its practical application in industrial contexts. To address this gap, we conducted a semistructured interview study with 13 industry practitioners to explore the current state of RAG adoption in real-world settings. Our study investigates how companies apply RAG in practice, providing (1) an overview of industry use cases, (2) a consolidated list of system requirements, (3) key challenges and lessons learned from practical experiences, and (4) an analysis of current industry evaluation methods. Our main findings show that current RAG applications are mostly limited to domain-specific QA tasks, with systems still in prototype stages; industry requirements focus primarily on data protection, security, and quality, while issues such as ethics, bias, and scalability receive less attention; data preprocessing remains a key challenge, and system evaluation is predominantly conducted by humans rather than automated methods.

Artificial intelligence risks are multidimensional in nature, as the same risk scenarios may have legal, operational, and financial risk dimensions. With the emergence of new AI regulations, the state of the art of artificial intelligence risk management seems to be highly immature due to upcoming AI regulations. Despite the appearance of several methodologies and generic criteria, it is rare to find guidelines with real implementation value, considering that the most important issue is customizing artificial intelligence risk metrics and risk models for specific AI risk scenarios. Furthermore, the financial departments, legal departments and Government Risk Compliance teams seem to remain unaware of many technical aspects of AI systems, in which data scientists and AI engineers emerge as the most appropriate implementers. It is crucial to decompose the problem of artificial intelligence risk in several dimensions: data protection, fairness, accuracy, robustness, and information security. Consequently, the main task is developing adequate metrics and risk models that manage to reduce uncertainty for decision-making in order to take informed decisions concerning the risk management of AI systems. The purpose of this paper is to orientate AI stakeholders about the depths of AI risk management. Although it is not extremely technical, it requires a basic knowledge of risk management, quantifying uncertainty, the FAIR model, machine learning, large language models and AI context engineering. The examples presented pretend to be very basic and understandable, providing simple ideas that can be developed regarding specific AI customized environments. There are many issues to solve in AI risk management, and this paper will present a holistic overview of the inter-dependencies of AI risks, and how to model them together, within risk scenarios.

It is known that big data analytics and AI pose a threat to privacy, and that some of this is due to some kind of "black box problem" in AI. I explain how this becomes a problem in the context of justification for judgments and actions. Furthermore, I suggest distinguishing three kinds of opacity: 1) the subjects do not know what the system does ("shallow opacity"), 2) the analysts do not know what the system does ("standard black box opacity"), or 3) the analysts cannot possibly know what the system might do ("deep opacity"). If the agents, data subjects as well as analytics experts, operate under opacity, then these agents cannot provide justifications for judgments that are necessary to protect privacy, e.g., they cannot give "informed consent", or guarantee "anonymity". It follows from these points that agents in big data analytics and AI often cannot make the judgments needed to protect privacy. So I conclude that big data analytics makes the privacy problems worse and the remedies less effective. As a positive note, I provide a brief outlook on technical ways to handle this situation.

The (generative) artificial intelligence (AI) era has profoundly reshaped the meaning and value of data. No longer confined to static content, data now permeates every stage of the AI lifecycle from the training samples that shape model parameters to the prompts and outputs that drive real-world model deployment. This shift renders traditional notions of data protection insufficient, while the boundaries of what needs safeguarding remain poorly defined. Failing to safeguard data in AI systems can inflict societal and individual, underscoring the urgent need to clearly delineate the scope of and rigorously enforce data protection. In this perspective, we propose a four-level taxonomy, including non-usability, privacy preservation, traceability, and deletability, that captures the diverse protection needs arising in modern (generative) AI models and systems. Our framework offers a structured understanding of the trade-offs between data utility and control, spanning the entire AI pipeline, including training datasets, model weights, system prompts, and AI-generated content. We analyze representative technical approaches at each level and reveal regulatory blind spots that leave critical assets exposed. By offering a structured lens to align future AI technologies and governance with trustworthy data practices, we underscore the urgency of rethinking data protection for modern AI techniques and provide timely guidance for developers, researchers, and regulators alike.

The evolution of generative AI systems exposes the challenges of traditional legal and ethical frameworks built around consent. This chapter examines how the conventional notion of consent, while fundamental to data protection and privacy rights, proves insufficient in addressing the implications of AI-generated content derived from personal data. Through legal and ethical analysis, we show that while individuals can consent to the initial use of their data for AI training, they cannot meaningfully consent to the numerous potential outputs their data might enable or the extent to which the output is used or distributed. We identify three fundamental challenges: the scope problem, the temporality problem, and the autonomy trap, which collectively create what we term a "consent gap" in AI systems and their surrounding ecosystem. We argue that current legal frameworks inadequately address these emerging challenges, particularly regarding individual autonomy, identity rights, and social responsibility, especially in cases where AI-generated content creates new forms of personal representation beyond the scope of the original consent. By examining how these consent limitations intersect with broader principles of responsible AI - including fairness, transparency, accountability, and autonomy - we demonstrate the need to evolve ethical and legal approaches to consent.

Figure 1 presents examples of NotebookLM's shortcomings Importantly, using NotebookLM to educate medical professionals presently risks of misleading them, as NotebookLM's lack Inaccurate responses given by NotebookLM to user queries; output is stylized for visual clarity. NotebookLM advises the user to tell their patients that eating rocks is healthy, citing the user's document. Passages from Dihan et al. advocating for use of NotebookLM (Column 1) which are associated with clinical and/or ethical concerns "Though NotebookLM is a commercial entity that does not abide by patient privacy regulations, it does represent an " A podcast generator can improve the way Given any set of documents, and especially those containing complex documents, LLMs may misinterpret and subsequently misrepresent some of their contents. "Rather than requiring active visual engagement through reading, podcasts allow NotebookLM can neither identify misinformation contained within uploaded files nor incorporate relevant information beyond the uploaded content. "[NotebookLM's] citations are automatically generated for all content that NotebookLM pulls from within these materials, No funding was received for the publication of this article.

Numerical reasoning over documents, which demands both contextual understanding and logical inference, is challenging for low-capacity local models deployed on computation-constrained devices. Although such complex reasoning queries could be routed to powerful remote models like GPT-4, exposing local data raises significant data leakage concerns. Existing mitigation methods generate problem descriptions or examples for remote assistance. However, the inherent complexity of numerical reasoning hinders the local model from generating logically equivalent queries and accurately inferring answers with remote guidance. In this paper, we present a model collaboration framework with two key innovations: (1) a context-aware synthesis strategy that shifts the query domains while preserving logical consistency; and (2) a tool-based answer reconstruction approach that reuses the remote-generated problem-solving pattern with code snippets. Experimental results demonstrate that our method achieves better reasoning accuracy than solely using local models while providing stronger data protection than fully relying on remote models. Furthermore, our method improves accuracy by 16.2% - 43.6% while reducing data leakage by 2.3% - 44.6% compared to existing data protection approaches.

This results in significant privacy risks, such as automated influence [7], manipulation [54], and potential security breaches. Yet, while companies invest heavily in acquiring and analyzing their users' personal data, users without extensive research or background knowledge lack awareness of the associated privacy risks [29] or have distorted perceptions of risks [31], which results in irrational decisions [1]. Regulations such as the GDPR [25] force companies to communicate data management practices and users' rights regarding their data in privacy policies, to enhance users' decision-making. However, evidence shows that companies focus on compliance, effectively targeting lawyers instead of users [78], so users rarely read privacy policies [61]. Using LLMs to automatically assess privacy policies is a promising approach to solve this issue [37, 72, 91]. Yet, no prior work evaluates their impact on understandability and risk awareness from a user's perspective through a user study. Additionally, to the best of our knowledge, no existing tool combines LLM-based automatic privacy policy assessment with: (i) dynamic evaluation criteria not focused on compliance but tailored to type of platform (e.g., e-commerce or health services); (ii) an interactive dashboard; and (iii) a chat for open conversations with the LLM with (iv) customizable explanations and responses that adapt to the user's preferences for detail and complexity. To address these gaps, we introduce PRISMe (Privacy Risk Information Scanner for Me), a Chrome extenstion with the above features designed to empower users in making informed privacy decisions. To evaluate PRISMe, we conducted a scenario-based, mixed-methods user study with a qualitative focus.

Context. The security of critical infrastructure has been a pressing concern since the advent of computers and has become even more critical in today's era of cyber warfare. Protecting mission-critical systems (MCSs), essential for national security, requires swift and robust governance, yet recent events reveal the increasing difficulty of meeting these challenges. Aim. Building on prior research showcasing the potential of Generative AI (GAI), such as Large Language Models, in enhancing risk analysis, we aim to explore practitioners' views on integrating GAI into the governance of IT MCSs. Our goal is to provide actionable insights and recommendations for stakeholders, including researchers, practitioners, and policymakers. Method. We designed a survey to collect practical experiences, concerns, and expectations of practitioners who develop and implement security solutions in the context of MCSs. Conclusions and Future Works. Our findings highlight that the safe use of LLMs in MCS governance requires interdisciplinary collaboration. Researchers should focus on designing regulation-oriented models and focus on accountability; practitioners emphasize data protection and transparency, while policymakers must establish a unified AI framework with global benchmarks to ensure ethical and secure LLMs-based MCS governance.

Availability attacks, or unlearnable examples, are defensive techniques that allow data owners to modify their datasets in ways that prevent unauthorized machine learning models from learning effectively while maintaining the data's intended functionality. It has led to the release of popular black-box tools for users to upload personal data and receive protected counterparts. In this work, we show such black-box protections can be substantially bypassed if a small set of unprotected in-distribution data is available. Specifically, an adversary can (1) easily acquire (unprotected, protected) pairs by querying the black-box protections with the unprotected dataset; and (2) train a diffusion bridge model to build a mapping. This mapping, termed BridgePure, can effectively remove the protection from any previously unseen data within the same distribution. Under this threat model, our method demonstrates superior purification performance on classification and style mimicry tasks, exposing critical vulnerabilities in black-box data protection.