arXiv:2602.04616v1 Announce Type: cross Abstract: As the paradigm of Human-Centered AI (HCAI) gains prominence, its benefits to society are accompanied by significant ethical concerns, one of which is the protection of individual privacy. This chapter provides a comprehensive overview of privacy within HCAI, proposing a human-centered privacy (HCP) framework, providing integrated solution from technology, ethics, and human factors perspectives. The chapter begins by mapping privacy risks across each stage of AI development lifecycle, from data collection to deployment and reuse, highlighting the impact of privacy risks on the entire system. The chapter then introduces privacy-preserving techniques such as federated learning and dif erential privacy. Subsequent chapters integrate the crucial user perspective by examining mental models, alongside the evolving regulatory and ethical landscapes as well as privacy governance. Next, advice on design guidelines is provided based on the human-centered privacy framework. After that, we introduce practical case studies across diverse fields. Finally, the chapter discusses persistent open challenges and future research directions, concluding that a multidisciplinary approach, merging technical, design, policy, and ethical expertise, is essential to successfully embed privacy into the core of HCAI, thereby ensuring these technologies advance in a manner that respects and ensures human autonomy, trust and dignity.

arXiv:2602.04678v1 Announce Type: cross Abstract: Time series forecasting in real-world applications requires both high predictive accuracy and interpretable uncertainty quantification. Traditional point prediction methods often fail to capture the inherent uncertainty in time series data, while existing probabilistic approaches struggle to balance computational efficiency with interpretability. We propose a novel Multi-Expert Learning Distributional Labels (LDL) framework that addresses these challenges through mixture-of-experts architectures with distributional learning capabilities. Our approach introduces two complementary methods: (1) Multi-Expert LDL, which employs multiple experts with different learned parameters to capture diverse temporal patterns, and (2) Pattern-Aware LDL-MoE, which explicitly decomposes time series into interpretable components (trend, seasonality, changepoints, volatility) through specialized sub-experts. Both frameworks extend traditional point prediction to distributional learning, enabling rich uncertainty quantification through Maximum Mean Discrepancy (MMD). We evaluate our methods on aggregated sales data derived from the M5 dataset, demonstrating superior performance compared to baseline approaches. The continuous Multi-Expert LDL achieves the best overall performance, while the Pattern-Aware LDL-MoE provides enhanced interpretability through component-wise analysis. Our frameworks successfully balance predictive accuracy with interpretability, making them suitable for real-world forecasting applications where both performance and actionable insights are crucial.

arXiv:2602.04735v1 Announce Type: cross Abstract: Large Language Models (LLMs) can acquire unintended biases from seemingly benign training data even without explicit cues or malicious content. Existing methods struggle to detect such risks before fine-tuning, making post hoc evaluation costly and inefficient. To address this challenge, we introduce Data2Behavior, a new task for predicting unintended model behaviors prior to training. We also propose Manipulating Data Features (MDF), a lightweight approach that summarizes candidate data through their mean representations and injects them into the forward pass of a base model, allowing latent statistical signals in the data to shape model activations and reveal potential biases and safety risks without updating any parameters. MDF achieves reliable prediction while consuming only about 20% of the GPU resources required for fine-tuning. Experiments on Qwen3-14B, Qwen2.5-32B-Instruct, and Gemma-3-12b-it confirm that MDF can anticipate unintended behaviors and provide insight into pre-training vulnerabilities.

arXiv:2602.04881v1 Announce Type: cross Abstract: Internet of Things (IoT) deployments operate in nonstationary, dynamic environments where factors such as sensor drift, evolving user behavior, and heterogeneous user privacy requirements can affect application utility. Continual learning (CL) addresses this by adapting models over time without catastrophic forgetting. Meanwhile, contrastive learning has emerged as a powerful representation-learning paradigm that improves robustness and sample efficiency in a self-supervised manner. This paper reviews the usage of \emph{contrastive continual learning} (CCL) for IoT, connecting algorithmic design (replay, regularization, distillation, prompts) with IoT system realities (TinyML constraints, intermittent connectivity, privacy). We present a unifying problem formulation, derive common objectives that blend contrastive and distillation losses, propose an IoT-oriented reference architecture for on-device, edge, and cloud-based CCL, and provide guidance on evaluation protocols and metrics. Finally, we highlight open unique challenges with respect to the IoT domain, such as spanning tabular and streaming IoT data, concept drift, federated settings, and energy-aware training.

arXiv:2505.19414v2 Announce Type: replace Abstract: The revolution in artificial intelligence (AI) has brought sustainable challenges in data center management due to the high carbon emissions and short cooling response time associated with high-power density racks. While machine learning (ML) offers promise for intelligent management, its adoption is hindered by safety and reliability concerns. To address this, we propose a multiphysics-informed machine learning (MPIML) framework that integrates physical priors into data-driven models for enhanced accuracy and safety. We introduce an integrated system architecture comprising three core engines: DCLib for versatile facility modeling, DCTwin for high-fidelity multiphysics simulation, and DCBrain for decision-making optimization. This system enables critical predictive and prescriptive applications, such as carbon-aware IT provisioning, safety-aware intelligent cooling control and battery health forecasting. An illustrative example on an industry-grade data center cooling control demonstrates that our MPIML approach reduces annual carbon emissions up to 200 kilotons compared with conventional methods while ensuring operational constraints are met. We conclude by outlining key challenges and future directions for developing autonomous and sustainable data centers.

arXiv:2503.01733v3 Announce Type: replace-cross Abstract: Smart homes equipped with ambient sensors offer a transformative approach to continuous health monitoring and assisted living. Traditional research in this domain primarily focuses on Human Activity Recognition (HAR), which relies on mapping sensor data to a closed set of predefined activity labels. However, the fixed granularity of these labels often constrains their practical utility, failing to capture the subtle, household-specific nuances essential, for example, for tracking individual health over time. To address this, we propose DISCOVER, a framework for discovering and annotating Patterns of Daily Living (PDL) - fine-grained, recurring sequences of sensor events that emerge directly from a resident's unique routines. DISCOVER utilizes a self-supervised feature extraction and representation-aware clustering pipeline, supported by a custom visualization interface that enables experts to interpret and label discovered patterns with minimal effort. Our evaluation across multiple smart-home environments demonstrates that DISCOVER identifies cohesive behavioral clusters with high inter-rater agreement while achieving classification performance comparable to fully-supervised baselines using only 0.01% of the labels. Beyond reducing annotation overhead, DISCOVER establishes a foundation for longitudinal analysis. By grounding behavior in a resident's specific environment rather than rigid semantic categories, our framework facilitates the observation of within-person habitual drift. This capability positions the system as a potential tool for identifying subtle behavioral indicators associated with early-stage cognitive decline in future longitudinal studies.

The previous article in this series, “Rules fail at the prompt, succeed at the boundary,” focused on the first AI-orchestrated espionage campaign and the failure of prompt-level control. This article is the prescription. The question every CEO is now getting from their board is some version of: What do we do about agent risk?

Across recent AI security guidance from standards bodies, regulators, and major providers, a simple idea keeps repeating: treat agents like powerful, semi-autonomous users, and enforce rules at the boundaries where they touch identity, tools, data, and outputs.

The following is an actionable eight-step plan one can ask teams to implement and report against:  

Constrain capabilities

These steps help define identity and limit capabilities.

1. Identity and scope: Make agents real users with narrow jobs

Today, agents run under vague, over-privileged service identities. The fix is straightforward: treat each agent as a non-human principal with the same discipline applied to employees.

Every agent should run as the requesting user in the correct tenant, with permissions constrained to that user’s role and geography. Prohibit cross-tenant on-behalf-of shortcuts. Anything high-impact should require explicit human approval with a recorded rationale. That is how Google’s Secure AI Framework (SAIF) and NIST AI’s access-control guidance are meant to be applied in practice.

The CEO question: Can we show, today, a list of our agents and exactly what each is allowed to do?

2. Tooling control: Pin, approve, and bound what agents can use

The Anthropic espionage framework worked because the attackers could wire Claude into a flexible suite of tools (e.g., scanners, exploit frameworks, data parsers) through Model Context Protocol, and those tools weren’t pinned or policy-gated.

The defense is to treat toolchains like a supply chain:

• Pin versions of remote tool servers.
• Require approvals for adding new tools, scopes, or data sources.
• Forbid automatic tool-chaining unless a policy explicitly allows it.

This is exactly what OWASP flags under excessive agency and what it recommends protecting against. Under the EU AI Act, designing for such cyber-resilience and misuse resistance is part of the Article 15 obligation to ensure robustness and cybersecurity.

The CEO question: Who signs off when an agent gains a new tool or a broader scope? How does one know?

3. Permissions by design: Bind tools to tasks, not to models

A common anti-pattern is to give the model a long-lived credential and hope prompts keep it polite. SAIF and NIST argue the opposite: credentials and scopes should be bound to tools and tasks, rotated regularly, and auditable. Agents then request narrowly scoped capabilities through those tools.

In practice, that looks like: “finance-ops-agent may read, but not write, certain ledgers without CFO approval.”

The CEO question: Can we revoke a specific capability from an agent without re-architecting the whole system?

Control data and behavior

These steps gate inputs, outputs, and constrain behavior.

4. Inputs, memory, and RAG: Treat external content as hostile until proven otherwise

Most agent incidents start with sneaky data: a poisoned web page, PDF, email, or repository that smuggles adversarial instructions into the system. OWASP’s prompt-injection cheat sheet and OpenAI’s own guidance both insist on strict separation of system instructions from user content and on treating unvetted retrieval sources as untrusted.

Operationally, gate before anything enters retrieval or long-term memory: new sources are reviewed, tagged, and onboarded; persistent memory is disabled when untrusted context is present; provenance is attached to each chunk.

The CEO question: Can we enumerate every external content source our agents learn from, and who approved them?

5. Output handling and rendering: Nothing executes “just because the model said so”

In the Anthropic case, AI-generated exploit code and credential dumps flowed straight into action. Any output that can cause a side effect needs a validator between the agent and the real world. OWASP’s insecure output handling category is explicit on this point, as are browser security best practices around origin boundaries.

The CEO question: Where, in our architecture, are agent outputs assessed before they run or ship to customers?

6. Data privacy at runtime: Protect the data first, then the model

Protect the data such that there is nothing dangerous to reveal by default. NIST and SAIF both lean toward “secure-by-default” designs where sensitive values are tokenized or masked and only re-hydrated for authorized users and use cases.

In agentic systems, that means policy-controlled detokenization at the output boundary and logging every reveal. If an agent is fully compromised, the blast radius is bounded by what the policy lets it see.

This is where the AI stack intersects not just with the EU AI Act but with GDPR and sector-specific regimes. The EU AI Act expects providers and deployers to manage AI-specific risk; runtime tokenization and policy-gated reveal are strong evidence that one is actively controlling those risks in production.

The CEO question: When our agents touch regulated data, is that protection enforced by architecture or by promises?

Prove governance and resilience

For the final steps, it’s important to show controls work and keep working.

7. Continuous evaluation: Don’t ship a one-time test, ship a test harness

Anthropic’s research about sleeper agents should eliminate all fantasies about single test dreams and show how critical continuous evaluation is. This means instrumenting agents with deep observability, regularly red teaming with adversarial test suites, and backing everything with robust logging and evidence, so failures become both regression tests and enforceable policy updates.

The CEO question: Who works to break our agents every week, and how do their findings change policy?

 8. Governance, inventory, and audit: Keep score in one place

AI security frameworks emphasize inventory and evidence: enterprises must know which models, prompts, tools, datasets, and vector stores they have, who owns them, and what decisions were taken about risk.

For agents, that means a living catalog and unified logs:

• Which agents exist, on which platforms
• What scopes, tools, and data each is allowed
• Every approval, detokenization, and high-impact action, with who approved it and when

The CEO question: If asked how an agent made a specific decision, could we reconstruct the chain?

And don’t forget the system-level threat model: assume the threat actor GTG-1002 is already in your enterprise. To complete enterprise preparedness, zoom out and consider the MITRE ATLAS product, which exists precisely because adversaries attack systems, not models. Anthropic provides a case study of a state-based threat actor (GTG-1002) doing exactly that with an agentic framework.

Taken together, these controls do not make agents magically safe. They do something more familiar and more reliable: they put AI, its access, and actions back inside the same security frame used for any powerful user or system.

For boards and CEOs, the question is no longer “Do we have good AI guardrails?” It’s: Can we answer the CEO questions above with evidence, not assurances?

This content was produced by Protegrity. It was not written by MIT Technology Review’s editorial staff.

Nature, Published online: 04 February 2026; doi:10.1038/s41586-025-10087-x

A multiancestry phenome-wide analysis of copy number variants in the UK Biobank genomes increases power to detect genetic associations with complex traits across human populations.

Nature, Published online: 04 February 2026; doi:10.1038/d41586-026-00212-9

Continual automated data collection was once possible only for the world’s richest labs. Now, sensor systems are commonplace and can be accessed from an app.

Nature Medicine, Published online: 04 February 2026; doi:10.1038/d41591-026-00007-5

Ethical research must not exploit scarcity as an experimental variable.

Among the big players in technology, Cisco is one of the sector’s leaders that’s advancing operational deployments of AI internally to its own operations, and the tools it sells to its customers around the world. As a large company, its activities encompass many areas of the typical IT stack, including infrastructure, services, security, and the design of entire enterprise-scale networks.

Cisco’s internal teams use a blend of machine learning and agentic AI to help them improve their own service delivery and personalise user experiences for its customers. It’s built a shared AI fabric built on patterns of compute and networking that are the product of years spent checking and validating its systems – battle-hardened solutions it then has the confidence to offer to customers. The infrastructure in play relies on high-performance GPUs, of course, but it’s not just raw horse-power. The detail is in the careful integration between compute and network stacks used in model training and the quite different demands from the ongoing load of inference.

Having made its name as the de facto supplier of networking infrastructure for the enterprise, it comes as no shock that it’s in network automation that some of its better-known uses of AI finds their place. Automated configuration workflows and identity management combine into access solutions that are focused on rapid network deployments generated by natural language.

For organisations looking to develop into the next generation of AI users, Cisco has been rolling out hardware and orchestration tools that are aimed explicitly to support AI workloads. A recent collaboration with chip giant NVIDIA led to the emergence of a new line of switches and the Nexus Hyperfabric line of AI network controllers. These aim to simplify the deployment of the complex clusters needed for top-end, high-performance artificial intelligence clusters.

Cisco’s Secure AI Factory framework with partners like NVIDIA and Run:ai is aimed at production-grade AI pipelines. It uses distributed orchestration, GPU utilisation governance, Kubernetes microservice optimisation, and storage, under the umbrella product description Intersight. For more local deployments, Cisco Unified Edge brings all the necessary elements – compute, networking, security, and storage – close to where data gets generated and processed.

In environments where latency metrics are critically important, AI processing at the edge is the answer. But Cisco’s approach is not necessarily to offer dedicated IIoT-specific solutions. Instead, it tries to extend the operational models typically found in a data centre and applies the same technology (if not the same exact methodology) to edge sites. It’s like data centre-grade security policies and configurations available to remote installations. Having the same precepts and standards in cloud and edge mean that Cisco accredited engineers can manage and maintain data centres or small edge deployments using the same skills, accreditation, knowledge, and experience.

Security and risk management figure prominently in the Cisco AI narrative. Its Integrated AI Security and Safety Framework applies high standards of safety and security throughout the life-cycle of AI systems. It considers adversarial threats, supply chain weakness, the risk profiles of multi-agent interactions, and multi-modal vulnerabilities as issues that have to be addressed regardless of the nature or size of any deployment.

Cisco’s work on operational AI also reflects broader ecosystem conversations. The company markets products for organisations wanting to make the transition from generative to agentic AI, where autonomous software agents carry out operational tasks. In most cases, this requires new tooling and new operational protocols.

Cisco’s future AI plans include continuing its central work in infrastructure provision for AI workloads. It’s also pursuing broader adoption of AI-ready networks, including next-gen wireless and unified management systems that will control systems across campus, branch, and cloud environments. The company is also expanding its software and platform investments, including its most recent acquisition (NeuralFabric), to help it build a more comprehensive software stack and product portfolio.

In summary, Cisco’s AI deployment strategy combines hardware, software, and service elements that embed AI into operations, giving organisations a route to production-grade systems. Its work can be found in large-scale infrastructure, systems for unified management, risk mitigation, and anywhere that connects distributed, cloud, and edge computing.

(Image source: Pixabay)

 

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.

The post How Cisco builds smart systems for the AI era appeared first on AI News.

arXiv:2602.02629v1 Announce Type: cross Abstract: The digitization of healthcare has generated massive volumes of Electronic Health Records (EHRs), offering unprecedented opportunities for training Artificial Intelligence (AI) models. However, stringent privacy regulations such as GDPR and HIPAA have created data silos that prevent centralized training. Federated Learning (FL) has emerged as a promising solution that enables collaborative model training without sharing raw patient data. Despite its potential, FL remains vulnerable to poisoning and Sybil attacks, in which malicious participants corrupt the global model or infiltrate the network using fake identities. While recent approaches integrate Blockchain technology for auditability, they predominantly rely on probabilistic reputation systems rather than robust cryptographic identity verification. This paper proposes a Trustworthy Blockchain-based Federated Learning (TBFL) framework integrating Self-Sovereign Identity (SSI) standards. By leveraging Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs), our architecture ensures only authenticated healthcare entities contribute to the global model. Through comprehensive evaluation using the MIMIC-IV dataset, we demonstrate that anchoring trust in cryptographic identity verification rather than behavioral patterns significantly mitigates security risks while maintaining clinical utility. Our results show the framework successfully neutralizes 100% of Sybil attacks, achieves robust predictive performance (AUC = 0.954, Recall = 0.890), and introduces negligible computational overhead (

arXiv:2602.03154v1 Announce Type: cross Abstract: Front-end personalization has traditionally relied on static designs or rule-based adaptations, which fail to fully capture user behavior patterns. This paper presents an AI driven approach for dynamic front-end personalization, where UI layouts, content, and features adapt in real-time based on predicted user behavior. We propose three strategies: dynamic layout adaptation using user path prediction, content prioritization through reinforcement learning, and a comparative analysis of AI-driven vs. rule-based personalization. Technical implementation details, algorithms, system architecture, and evaluation methods are provided to illustrate feasibility and performance gains.

This AI doctor is licensed in all 50 states, the startup says. The deal was led by CRV and Kleiner Perkins.

When is an AI doctor a medical device?

Call it a sign of things to come. A startup called Doctronic made a splash recently when it announced the use AI to renew prescriptions without clinician input in the state of Utah. Something didn’t sit right with me about the announcement. Sure it got approval from Utah, but why isn’t it a medical device subject to Food and Drug Administration review? The company claimed it was “the practice of medicine” and so exempt from FDA authority. That didn’t seem entirely right either. 

So I did some asking around and after talking to over a dozen executives, legal scholars, and policy experts, it turns out the question is not nearly as clear-cut as Doctronic would have us believe. Indeed, it appears the company may be planning to market a medical device without authorization. In my story, I explain the law and why it all matters.

Read more here

Continue to STAT+ to read the full story…

© Adobe

Utah’s recent announcement that it was partnering with a health tech startup that will use artificial intelligence to renew drug prescriptions may offer a glimpse of the futuristic version of AI medicine that’s long been foretold by technologists and venture capitalists.

But it’s only possible because of some pre-approved rule breaking — and may prove a broader test of the Food and Drug Administration’s authority to evaluate a new wave of clinical AI products, according to interviews with executives and experts.

In January, Utah regulators said they had signed an agreement with a startup called Doctronic to launch an AI system that will perform a clinical evaluation of patients and, when deemed appropriate, renew some 200 common medications autonomously.

Continue to STAT+ to read the full story…

© Illustration: Camille MacMillin/STAT; Photos: Adobe

npj Digital Medicine, Published online: 03 February 2026; doi:10.1038/s41746-026-02385-z

Digital intervention mylovia improves sexual functioning in women with sexual dysfunction in randomized controlled trial

Background: The transformation of digital health technologies has reshaped how healthcare is delivered, particularly in primary care. However, despite the advantages of these innovations, older adults remain among the most resistant users. Traditional technology adoption models may not fully capture the complexity of this reluctance, which is shaped not only by usability challenges but also by emotional, psychological, and identity-related concerns. Innovation Resistance Theory (IRT) offers a complementary framework focused on understanding barriers to adoption rather than solely on facilitators. Objective: To map and synthesize evidence on older adults’ resistance to digital health technologies in primary care through the lens of IRT, and to examine how empirically observed resistance factors align with, extend, or refine IRT’s functional and psychological barriers. Methods: A scoping review combined with concept-driven thematic synthesis was conducted. Empirical studies published between 2014 and 2025 were identified through systematic searches across five databases: PubMed, CINAHL, Ovid Medline, Web of Science, and Scopus. Inclusion criteria focused on studies examining barriers or resistance to digital health use among older adults aged 60 and above in primary care settings. The search was guided by terms related to “older adults”, “digital health”, “eHealth”, “telemedicine”, and “technology resistance”. After screening and reviewing the full texts, data were extracted into a structured matrix, and findings were organized according to the five dimensions of the IRT: usage, value, risk, tradition, and image barriers. Results: Of 4,976 identified records, seventeen studies met the inclusion criteria. Functional barriers included usability challenges, interface complexity, and age-related impairments. Psychological resistance was frequently linked to emotional discomfort, symbolic misalignment, and concerns about the loss of relational care. Value and risk concerns included distrust in diagnostics accuracy, concerns regarding privacy and data security, and skepticism about care quality. Traditional preferences for face-to-face interactions and generational digital divides further reinforced image-based resistance. A key finding was the interaction between barriers, where low self-efficacy and technology anxiety create feedback loops that reinforce avoidance behaviors. Conclusions: Resistance to digital health among older adults is not simply a lack of adoption but a complex, emotionally grounded process involving functional, psychological, and identity-based barriers. Interventions must go beyond technical usability to rebuild emotional trust, preserve the relational aspects of care, and align digital solutions with the values and expectations of older adults. Innovation Resistance Theory offers a comprehensive framework for understanding these multifaceted dynamics and serves as a valuable guide for policy development, user-centered design, and future research Clinical Trial: None