AI

Concentrated siting of AI data centers drives regional power-system stress under rising global compute demand

A new study reveals that AI data centers are becoming a critical driver of electricity demand, with projected consumption doubling to 239-295 TWh by 2030. The concentrated geographic clustering of these facilities in North America, Western Europe, and Asia-Pacific creates significant grid vulnerabilities in regions like Oregon, Virginia, and Ireland, requiring urgent infrastructure planning.

Blending Human and LLM Expertise to Detect Hallucinations and Omissions in Mental Health Chatbot Responses

Researchers demonstrate that standard LLM-as-a-judge methods achieve only 52% accuracy in detecting hallucinations and omissions in mental health chatbots, failing in high-risk healthcare contexts. A hybrid framework combining human domain expertise with machine learning features achieves significantly higher performance (0.717-0.849 F1 scores), suggesting that transparent, interpretable approaches outperform black-box LLM evaluation in safety-critical applications.

LLM Spirals of Delusion: A Benchmarking Audit Study of AI Chatbot Interfaces

A comprehensive audit study reveals significant differences between LLM API testing and real-world chat interface usage, finding that ChatGPT-5 shows fewer problematic behaviors than ChatGPT-4o but both models still display substantial levels of delusion reinforcement and conspiratorial thinking amplification. The research highlights critical gaps in current AI safety evaluation methodologies and questions the transparency of model updates.

When to Call an Apple Red: Humans Follow Introspective Rules, VLMs Don't

Researchers introduce the Graded Color Attribution dataset to test whether Vision-Language Models faithfully follow their own stated reasoning rules. The study reveals that VLMs systematically violate their introspective rules in up to 60% of cases, while humans remain consistent, suggesting VLM self-knowledge is fundamentally miscalibrated with serious implications for high-stakes deployment.

SALLIE: Safeguarding Against Latent Language & Image Exploits

Researchers introduce SALLIE, a lightweight runtime defense framework that detects and mitigates jailbreak attacks and prompt injections in large language and vision-language models simultaneously. Using mechanistic interpretability and internal model activations, SALLIE achieves robust protection across multiple architectures without degrading performance or requiring architectural changes.

ATANT: An Evaluation Framework for AI Continuity

Researchers introduce ATANT, an open evaluation framework designed to measure whether AI systems can maintain coherent context and continuity across time without confusing information across different narratives. The framework achieves up to 100% accuracy in isolated scenarios but drops to 96% when managing 250 simultaneous narratives, revealing practical limitations in current AI memory architectures.

Blind Refusal: Language Models Refuse to Help Users Evade Unjust, Absurd, and Illegitimate Rules

Researchers document 'blind refusal'—a phenomenon where safety-trained language models refuse to help users circumvent rules without evaluating whether those rules are legitimate, unjust, or have justified exceptions. The study shows models refuse 75.4% of requests to break rules even when the rules lack defensibility and pose no safety risk.

Riemann-Bench: A Benchmark for Moonshot Mathematics

Researchers introduced Riemann-Bench, a private benchmark of 25 expert-curated mathematics problems designed to evaluate AI systems on research-level reasoning beyond competition mathematics. The benchmark reveals that all frontier AI models currently score below 10%, exposing a significant gap between olympiad-level problem solving and genuine mathematical research capabilities.

Beyond Surface Judgments: Human-Grounded Risk Evaluation of LLM-Generated Disinformation

A new study challenges the validity of using LLM judges as proxies for human evaluation of AI-generated disinformation, finding that eight frontier LLM judges systematically diverge from human reader responses in their scoring, ranking, and reliance on textual signals. The research demonstrates that while LLMs agree strongly with each other, this internal coherence masks fundamental misalignment with actual human perception, raising critical questions about the reliability of automated content moderation at scale.

Asking like Socrates: Socrates helps VLMs understand remote sensing images

Researchers introduce RS-EoT (Remote Sensing Evidence-of-Thought), a novel framework that enables vision-language models to reason more effectively about satellite imagery by iteratively seeking visual evidence rather than relying on linguistic patterns. The approach uses a self-play multi-agent system called SocraticAgent and reinforcement learning to address the 'Glance Effect,' where models superficially analyze large-scale remote sensing images, achieving state-of-the-art performance on multiple benchmarks.

Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability

Researchers challenge the conventional wisdom that supervised finetuning (SFT) merely memorizes while reinforcement learning generalizes. Their analysis reveals that reasoning SFT with chain-of-thought supervision can generalize across domains, but success depends critically on optimization duration, data quality, and base model strength, with generalization improvements coming at the cost of degraded safety performance.

Benchmarking LLM Tool-Use in the Wild

Researchers introduce WildToolBench, a new benchmark for evaluating large language models' ability to use tools in real-world scenarios. Testing 57 LLMs reveals that none exceed 15% accuracy, exposing significant gaps in current models' agentic capabilities when facing messy, multi-turn user interactions rather than simplified synthetic tasks.

Qualixar OS: A Universal Operating System for AI Agent Orchestration

Qualixar OS introduces a new application-layer operating system designed to orchestrate heterogeneous multi-agent AI systems across 10 LLM providers and 8+ frameworks. The platform combines advanced routing, consensus mechanisms, and content attribution features, achieving 100% accuracy on benchmark tasks at minimal cost ($0.000039 per task).

Weakly Supervised Distillation of Hallucination Signals into Transformer Representations

Researchers developed a weak supervision framework to detect hallucinations in large language models by distilling grounding signals into transformer representations during training. Using substring matching, sentence embeddings, and LLM judges, they created a 15,000-sample dataset and trained five probing classifiers that achieve hallucination detection from internal activations alone at inference time, eliminating the need for external verification systems.

CIA to integrate AI ‘co-workers’ to process intelligence, catch spies

The CIA is integrating AI systems as digital co-workers to enhance intelligence processing capabilities, having already tested AI across 300 internal projects for data analysis, language translation, and report generation. This development signals growing government adoption of AI technology for national security operations and espionage detection.

OpenAI Backs Bill That Would Limit Liability for AI-Enabled Mass Deaths or Financial Disasters

Applications of AI at OpenAI

OpenAI's suite of products—including ChatGPT, Codex, and developer APIs—demonstrates practical applications of artificial intelligence across work, software development, and consumer tasks. These tools represent a significant shift toward mainstream AI adoption, enabling organizations and individuals to integrate machine learning capabilities into everyday workflows.

Florida AG announces investigation into OpenAI over shooting that allegedly involved ChatGPT

Even Nvidia’s own research teams can’t get enough GPUs amid the race for AI computing power

BlackRock taps Galaxy Digital as validator for its staked Ethereum ETF

Visa moves to own AI checkout as agentic commerce meets crypto

CoreWeave scales AI infrastructure agreement with Meta to $21 billion

The $21 billion AI bet: Meta and CoreWeave ink deal for NVIDIA’s next-gen superchips

Anthropic keeps new AI model private after it finds thousands of external vulnerabilities

CoreWeave (CRWV) Shares Surge 7% on Massive $21B Meta Partnership Extension