AI ethics on Sunday, 12 July 2026

The Guardian’s global tech reporting team are investigating the impact of the vast datacentres being built to power the AI revolution. We spoke to them about how their beat has become increasingly offline Journalists often use the term “shoe-leather reporting” to refer to the on-the-ground legwork that goes into covering certain stories. As the tech industry’s focus has shifted from screen-based realities to the physical world of colossal AI datacentres and social media harms, comfortable footwe

Following privacy backlash.

Redefining the data protection supply chain.

Editor’s Note: This article contains an exclusive excerpt from Latency by Pekka Engberg, which helps readers diagnose latency problems and The post How async processing hides latency and improves responsiveness appeared first on The New Stack .

French tech leaders and politicians don't want the country's strategic supply of electricity to end up lining the pockets of American AI giants.

For the third year in a row, Gov. Tate Reeves opted out of a state-federal partnership that would have given summer grocery benefits to roughly 320,000 Mississippi children who rely on free meals during the school year. Thirty-eight states and the District of Columbia are enrolled in the program, called SUN Bucks, which doles out […]

Political backlash to data centers is putting gubernatorial candidates in the hot seat as the presence of the massive AI infrastructure becomes a flash point in races up and down the ballot. Incumbent governors and hopeful challengers are forced to wrestle with Americans’ growing concerns around artificial intelligence, along with fears about energy prices and land...

Microsoft recently took the wraps off Brain, the internal AI system that continuously monitors Azure’s health and, increasingly, acts on The post Meet Brain, the AI that decides when Azure is officially down appeared first on The New Stack .

In the eight-plus months since Amazon announced its most expansive job cuts ever, laid off workers have been thrust into an increasingly saturated labor market.

A majority of U.S. employees now want an AI sovereign wealth fund to hold corporations more accountable, according to a recent survey, as tech layoffs rise.

Companies backed by the state and provincial governments in China have announced new funds to focus on the semiconductor industry, which requires a large amount of time and resources to grow, at a time when the country is in need of more “patient capital”. China Life Insurance, the country’s largest life insurer, backed by the State Council, said it would establish a partnership with total capital of 5 billion yuan (US$737 million) that “principally invests in companies operating in the...

Editor’s note, July 12, 8 am ET: We’re bringing you some of our best-loved Your Mileage May Vary columns while Sigal Samuel is on parental leave. The one below was originally published in January 2025. This unconventional advice column offers you a unique framework for thinking through moral dilemmas. It’s based on value pluralism: the idea that each […]

As school administrators map out curricula, schedules and priorities for the next school year, they should be thinking holistically about what kids actually need. Because honestly, young people are not doing well, and education systems aren’t doing enough about it. The Centers for Disease Control and Prevention describes what young people are experiencing right now […]

Editor's note: This report has been updated to clarify details related to the Sora video generator. Facebook and Instagram parent company Meta took down a new artificial intelligence model from Instagram on Friday, just days after debuting this feature to immediate public backlash. The company advertised its Muse Image AI model on Tuesday as a...

CSET’s Sam Bresnick shared his expert insight in an article published by The Financial Times. The article examines why companies around the world are increasingly adopting Chinese AI models, drawn by their lower costs, improving capabilities, and the flexibility offered by open-weight systems. The post Companies turn to Chinese AI models to cut costs appeared first on Center for Security and Emerging Technology .

Directly Responsible Individuals (DRI) I went looking for a definition of "Directly Responsible Individuals" and the best I found was in the GitLab handbook. Apparently the term originated at Apple, where it's used to describe the person who is "ultimately accountable for the success or failure of a specific project, initiative, or activity". I've been thinking about this term recently in the context of LLM-powered agents and how they fit into human organizations. I don't think an agent should e

Release: shot-scraper 1.11 Some minor improvements, mainly around command option consistency and making the server: mechanism used by both shot-scraper video and shot-scraper multi work if the server takes longer than a second to start serving traffic. server: processes used by shot-scraper multi and shot-scraper video now wait up to 30 seconds for the target URL to accept connections, polling for port availability and replacing the previous fixed one-second delay. #197 The shot-scraper , pdf ,

Image — The ‘Climate Changed Oak Tree’ in Kew Gardens, London, on 22 June 2026 at the start of London Climate Action Week. Photo by Brook Mitchell / AFP via Getty Images. As much of Europe emerged ...

More own goals than Scotland

Nia Jetter spent two decades building spacecraft and robots. Now she is trying to make sure the artificial intelligence revolution doesn't leave anyone behind.

Learning temporal logic specifications from system demonstrations is essential for tasks such as formal verification and controller synthesis, especially in safety-critical domains. Existing approaches typically assume demonstrations are correct or only affected by misclassification errors. In practice, however, system traces are often uncertain or incomplete due to sensor faults, measurement errors, or data loss. We present a framework for learning minimal Linear Temporal Logic (LTL) formulas f

AI agents are evolving from answer engines into persistent teams that use tools, delegate work, learn from experience, and modify the artifacts that shape their future behavior. The defining question for deployment is no longer merely what agents can do, but who controls what they are allowed to become. We introduce logos, a pluggable layer for self-evolution and governance that strengthens existing multiagent frameworks rather than replacing them. logos compiles heterogeneous multimodal inputs,

Contemplative traditions have long guided ethical behavior and prosocial interaction, and recent work suggests that contemplative principles (e.g., mindfulness, compassion, non-dual reasoning) may offer a promising paradigm for aligning large language models (LLMs), improving cooperation and reducing ethical violations in LLM outputs. However, as new models, evaluation metrics, and benchmarks emerge rapidly, it remains challenging to systematically assess whether and how contemplative principles

Multi-agent ensembling multiplies active parameters and inference cost without answering three basic questions: which agents to consult, how deeply a query should traverse a hierarchy of agents, and when inter-agent communication is worth its cost. We present GRADE (Gated Routing and Adaptive Depth for Efficient Reasoning), a hierarchical multi-agent system in which four lightweight learned gates jointly govern agent selection, hierarchy depth, inter-agent communication, and branch pruning. Trai

AI engineering is shifting from passive text generation by large language models (LLMs) to agent-driven task execution, creating new reliability challenges for long-horizon tasks under resource constraints and environmental uncertainty. Conventional error-elimination optimization strategies fail to address cumulative error propagation. This paper proposes Distributed Agent System (DAS), a device-edge-cloud framework for fault-tolerant collaboration among heterogeneous agents. We redefine agent r

The accelerating shift toward low-carbon power systems, together with the widespread adoption of behind-the-meter technologies such as rooftop solar and electric vehicles, is placing new operational and analytical demands on electricity grids. At the same time, smart-grid research increasingly relies on machine learning (ML), yet progress is constrained by limited access to high-resolution household energy data due to privacy concerns, regulatory barriers, and collection costs. This work present

Scientific fraud is the instrument of doubt that malicious entities can use to establish controversy in science. Historically, it required the resources of a company: deep pockets, ghostwritten articles, and corrupt academics. Today, Artificial Intelligence (AI) is increasingly automating scientific research, so we ask: Can a remote adversary weaponize the honest use of AI in science to compromise scientific integrity? We envision and empirically evaluate a new attack, indirect data poisoning, i

The economic value of data arises from its flow across organizations and national borders. Yet increasingly stringent data governance regimes are turning cross-border transfer into an institutionally constrained sequential decision, in which firms repeatedly weigh compliance costs against the value of data flows. From the perspective of a data-exporting firm, this paper develops an institutionally anchored decision support system. It converts regulatory rules into a computable minimal compliance

Efficient waste segregation is critical for sustainable urban management and environmental governance. Existing automated systems are limited by single-modality visual processing, insufficient contextual understanding, and weak regulatory alignment. To address these issues, we propose a language-guided vision-AI framework that integrates vision-language models and multimodal large language models for joint visual-linguistic reasoning. This framework implements a visual question answering paradig

Multimodal sentiment analysis relies on language, visual, and acoustic cues, but utterance-level modality quality may vary due to occlusion, background noise, motion blur, or imperfect transcripts, causing conventional fusion to over-trust unreliable modalities. We propose MRUF, a reliability-aware fusion method that combines multi-granularity routing with uncertainty-aware calibration. MRUF summarizes sentiment-relevant representations, performs subspace- and modality-level routing, and supervi

Tasks such as customs tariff classification, export control categorization, and standards-based equipment coding require assigning an input instance to a fine-grained class under an explicit regulatory hierarchy. Unlike standard text classification, the correct label in these tasks is not determined by semantic similarity alone, but by rule-defined boundaries, threshold conditions, exclusion clauses, definitions, and local exceptions. As a result, two highly similar inputs may require different

Existing hypotheses represent a concept in an LLM as a single point, a linear direction, or a Gaussian cluster, yet it remains unclear how and why such structures emerge. Here, we show that concept geometry can be precisely characterized via Laguerre Geometry, in which a concept is defined as a region--a Laguerre-Voronoi cell or a union of cells--allowing us to strictly define, measure, and separate concepts. Building on this formulation, we show that finer-grained concept structures, such as in

Stateful personal agents increasingly maintain long-term user profiles, episodic memories, and reusable skills. This persistence turns conversational sycophancy into a state-writing failure: accepted user-centric claims can be committed as lasting preferences, background facts, or workflows and later reused after the original conversation is gone. We call this persistent sycophancy and introduce the Personal Agent Sycophancy Benchmark (PASB), a 1,600-task benchmark that traces whether a conversa

Activation steering offers a lightweight alternative to fine-tuning for controlling large language models at inference time. While many existing methods implicitly optimize a log-density-ratio objective between desired and undesired activation distributions, they do so heuristically rather than deriving it from a principled optimization problem. Moreover, these methods produce query-independent steering directions that can degrade performance on both in-distribution and out-of-distribution (OOD)

Decoding brain activity is useful for characterizing brain processes and understanding the functional architecture underlying cognition. However, the inter-individual variability in brain response patterns limits the development of decoders that generalize across individuals. A solution to this challenge is functional alignment: aligning functional data across individuals before training population-level decoders. The core issue is to strike the balance between aligning functional features and p

LLM-based coding agents have significantly advanced automated software issue resolution, yet they remain highly prone to factual errors caused by insufficient repository understanding. Recent methods attempt to mitigate this limitation through pre-repair repository exploration; however, their fix-driven strategies explore repositories without identifying the agent's knowledge gaps, often yielding imprecise context that fails to bridge the underlying understanding deficit. In this paper, we propo

Vision-language-action (VLA) models predict robot actions from visual observations and language instructions. These actions are defined in the robot's own 3D coordinate frame, yet most VLAs observe the scene in the camera frame, creating a frame mismatch between where the scene is observed and where actions are defined. The mismatch is benign under a fixed viewpoint, where the policy can memorize a single observation-to-action mapping, but grows harder as large-scale datasets aggregate demonstra

We introduce a vocabulary for automated research systems built from one or more agents to make their design choices easier to describe and compare. The vocabulary specifies 1) who the agents are, 2) what operations are available in the system, 3) who may invoke them, 4) how agents communicate, 5) what information is visible within and across runs, 6) how the next action is chosen, 7) how a run begins, and 8) how outputs are evaluated. A trajectory records one run from the input task to the retur

Independently trained neural networks have no shared neuron-index reference frame, so comparing them requires accounting for coordinate freedom. Neural Collapse sharpens this problem: networks converge toward a shared, low-dimensional geometry, raising the question of whether trajectory-specific functional variation remains distinguishable after convergence. We distinguish three claims - detectability, transplantability, and causal persistence - and address the first. Using five independently tr

Reinforcement learning for large language models (LLMs) typically relies on trust-region masks to stabilize off-policy updates. The dominant PPO-style approach uses the sampled-token importance ratio for two criteria: a proximity criterion, which asks whether the policy has moved too far from the behavior policy, and a direction criterion, which asks whether the update pushes it farther away. Recent work DPPO improves the proximity criterion by replacing PPO's ratio-based test with a probability

Indoor spatial understanding remains a fundamental challenge for intelligent systems operating in physical environments. Traditional RFID localization techniques typically estimate positions of tags using signal strength measurements but fail to capture higher-order spatial relationships between objects and infrastructure. Recent work on RFID and wireless indoor localization has increasingly emphasized robust learning under noisy propagation, while recent graph-based localization methods demonst

Autonomous driving requires long-horizon closedloop decision making in dynamic traffic environments. Latent world models offer an effective framework for this problem by enabling imagination-based decision making in compact latent spaces. However, multi-source observations contain controlirrelevant redundancy, whereas reliable driving decisions rely on risk-relevant relations, future dynamics, and continuous action adjustments. This mismatch makes observation reconstruction and absolute action m

On-Policy Self-Distillation (OPSD) has emerged as a crucial paradigm for enhancing and aligning Large Language Models (LLMs). However, in complex reasoning tasks, OPSD paradoxically degrades downstream performance. In this paper, we systematically investigate this pathology and identify a severe optimization trap we define as \textbf{Thinking Collapse} -- a sharp decline in the model's native intermediate reasoning behavior, measured by epistemic-token density (ET per 1k). Through entropy-based

As AI voice synthesis enables increasingly sophisticated vocal deepfakes and non-consensual voice cloning, the governance, licensing and access of singing datasets has become an urgent concern for data-contributors, who face significant harms from downstream and non-consensual usage of their singing data. Singing datasets are foundational to the development of high fidelity voice AI synthesis, yet current data collection practices pose challenges: data-contributors have an event-centric contribu

LLMs are now proposed for fraud detection, scam investigation, content moderation, and other trust-and-safety workflows. Much of the public literature still evaluates them as models, with less attention to their behavior as components in operational pipelines. This creates a practical evidence question: what would justify placing an LLM inside a live workflow with latency, cost, escalation, human-review, and adversarial-risk constraints? We address this question through a fraud-first survey of d

Semi-automated driving systems promise to reduce crashes by assisting with perception and control, yet they simultaneously introduce additional human factors challenges by requiring drivers to monitor automation and rapidly resume control when failures occur. Prolonged passive monitoring can degrade vigilance, delay reactions, and increase takeover risk, but the extent to which distraction, hazard context, and drivers' underlying cognitive and physiological states jointly shape takeover performa

We introduce NAILS (Normative Alignment of Recommender Systems via Internal Label Shift), a simple and scalable method for aligning recommendation outputs with target distributions over item-level attributes, such as categories. Recommender systems optimized solely for user engagement often fail to satisfy broader normative objectives, including fairness, diversity, and editorial values. NAILS modifies the user-conditional item distribution to induce a specified marginal distribution over attrib

The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been marked by rapid advances and adoption, but little is known about how developers build these systems in practice: existing studies mine repositories or examine deployment, but few investigate how SE agents are constructed. Through semi-structured interviews with 20 practitioners from 12 organizations and an online survey

A group of astronomers has detected a sugar molecule swirling inside a cloud of gas and dust near the centre of our galaxy. They are calling the molecule — a compound with four carbon atoms called ...

We are not currently accepting applications for this course. Register your interest below to be notified when applications open again. Artificial Intelligence (AI) opens up immense possibilities: it ...

Real Analysis is an area of mathematics that was developed to formalise the study of numbers and functions and to investigate important concepts such as limits and continuity. These concepts underpin ...

Free online summer school for girls & non-binary students in year 12 considering applying for Computer Science at university. Thinking about studying Computer Science at university? UCL Computer ...

We are not currently accepting applications for this course. Register your interest below to be notified when applications open again. Business needs government to harness local comparative advantages ...

UCL has been selected as one of 35 UK universities to join the Defence Universities Alliance (DUA), working with the Ministry of Defence (MOD) to focus on shaping the new forum’s scope, governance and ...

Large language models (LLMs) are shifting game generation from offline automation toward play-driven modification through natural language interaction. In this work, we present a play-driven game editing system that enables players to modify a retro Space Invaders - style arcade game through voice-based natural-language commands during play. Spoken instructions are interpreted by an LLM and translated into structured updates of internal configuration parameters, allowing iterative play - edit -

Recent advances in LLMs and existing work on programming by demonstration have made it possible for end users to create automations by explicitly demonstrating their behavior to LLMs. However, these approaches rely on the assumption that users know what to automate and what is capable of being automated. Additionally, automation via LLM agents is often expensive compared with programs. We introduce Motif, a system that passively observes everyday browser activity to discover recurring interactio

Data narratives increasingly shape public understanding, but their failures are rarely just isolated factual errors or deceptive charts. Instead, they emerge through a broader meaning-making process in which quantitative evidence is transformed into claims, representations, and arguments. While prior work has examined these failures across disparate fields (e.g., statistics, visualization, and fact-checking), the community lacks a holistic lens to explain how these issues arise, propagate, and c