AI News Archive: May 5, 2026 — Part 15
Sourced from 500+ daily AI sources, scored by relevance.
- Multilingual Evaluation of a Large Language Model-Based Primary Care Chatbot
Pre-visit planning has the potential to reduce EHR documentation burden while improving workflow efficiency, care quality, and patient-provider engagement. Large language model (LLM) chatbots show promise for supporting this task, but while their English-centric development suggests a potential for disparity, the extent to which these concerns translate into performance degradation in multilingual clinical settings remains unclear. In this mixed-methods study, we systematically evaluate the multilingual capabilities of PCP-Bot, an English-developed LLM-based (GPT-4o) clinical chatbot that collects patient concerns and generates structured, physician-ready summaries (~200 words) under structured output constraints. We enrolled 31 bilingual individuals (11 Mandarin, 10 Spanish, 10 Hindi) to role-play as patients to evaluate the PCP-Bot, interacting with it across five synthetic clinical cases in both English and a second language. Participants completed a structured survey comprising bas
- Screening for Rheumatic Heart Disease in Asymptomatic Children using Machine Learning from Electrocardiograms
Early detection of Rheumatic Heart Disease (RHD) is essential in reducing its associated mortality and late complications. In resource-limited settings, automated detection using low-cost electrocardiogram (ECG) sensors can enhance prevention efforts. However, its effectiveness as a potential RHD screening tool in at-risk populations remains unexplored. This study aimed to investigate the utility of machine learning for classifying RHD in a cohort screened for RHD using low-cost ECG devices. The ECGs were collected from 611 at-risk schoolchildren using KardiaMobile, where 47 were confirmed RHD and 564 were healthy. First, the ECG fiducial points were annotated using a publicly available prominence-based delineator. Then, temporal, frequency, wavelet, and visibility graph-based features were extracted from six-leads and fed to the XGBoost classifier. A 10-fold cross-validation was used at different prediction score thresholds to obtain target sensitivity (Se) for screening RHD. Single-l
- Disentangling the contribution of disease genes to drug therapeutic and side effects
Most clinical trials fail due to either lack of efficacy or safety concerns. Human genetics can address both failure reasons: disease-associated genes are not only promising therapeutic targets but also predict drug side effects. However, because the same genetic signal underlies both outcomes, we need methods that disentangle which disease genes mediate therapeutic benefit versus adverse side effects. We use DraphNet, our previously developed model that maps drug molecular effects onto disease genes to generate two gene sets per drug: one linked to its therapeutic effects (IND genes) and one linked to its side effects (SE genes). We show that IND and SE genes overlap for 76% of the tested drugs (compared to a null model). We also show that drugs sharing greater IND similarity also have greater SE similarity ({rho}=0.57, p<1e-300). To show how our approach enables insights into drug biology, we construct groupings of drugs based on their IND and SE genes. We find that drugs in the same
- Detection of Hepatocellular Carcinoma from B-Mode and Contrast-Enhanced Ultrasound Using a Dual-Path Convolutional Network
Background: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide, with particularly severe consequences in sub-Saharan Africa where access to advanced diagnostic imaging remains limited. Ultrasound is the most widely available imaging modality in low-resource settings, yet its sensitivity for detecting early-stage HCC remains insufficient when used in conventional B-mode alone.Methods: We present a dual-path convolutional neural network (CNN) that jointly analyzes B-mode and contrast-enhanced ultrasound (CEUS) images for automated HCC detection. The model processes 1,057 labeled liver ultrasound images from 85 patients sourced from The Cancer Imaging Archive, a publicly available single-center dataset. A preprocessing pipeline extracts liver-centered regions of interest from heterogeneous DICOM files, including automatic separation of dual-panel B-mode and CEUS frames. Each imaging modality is processed through a dedicated ResNet-34 backbone initializ
- A data-driven Alzheimer's disease progression simulator for retrospective validation and prospective Phase III power design
Anti-amyloid immunotherapies have recently demonstrated the first significant slowing of cognitive decline in Alzheimer's disease (AD), yet clinical benefit varies markedly across drugs and scales with the completeness of amyloid clearance. Pharmacokinetic/pharmacodynamic (PK/PD) models are currently the standard tool for trial simulation, but they typically operate on single biomarkers and rely on drug-concentration assumptions, leaving the multi-scale cascade from amyloid clearance through tau, neurodegeneration, and cognition largely unmodelled. No existing framework has been jointly validated against the quantitative outcomes of multiple real-world phase III trials, spanning clearance kinetics, multi-modal biomarker trajectories, and statistical power. We present a trial simulation platform based on SimulAD, a disease progression model trained exclusively on longitudinal observational data from ADNI, with no access to trial-arm labels or drug-specific outcomes. SimulAD encodes inte
- Neural signatures of real-world turning during naturalistic locomotion in Parkinson's Disease
Turning is a complex motor behavior that frequently triggers freezing of gait and falls in Parkinson's disease (PD), yet its neural dynamics in naturalistic settings remain unknown. Using chronic at-home intracranial recordings in four subjects with PD, we show that turning is marked by premotor cortical beta desynchronization driven by reduced burst rate. These findings identify a robust signature of ecological turning and implicate beta dynamics in adaptive motor transitions.
- Comparative analyses of Alzheimers disease blood biomarkers and cognitive domains
INTRODUCTION: Whether Alzheimers disease (AD) blood biomarker-cognition associations differ across cognitive domains, analytic context, and biomarker modeling strategy in population-based cohorts is unclear. METHODS: In 1,170 older adults from the Health and Retirement Study Harmonized Cognitive Assessment Protocol, we examined cross-sectional (2016) and prospective (2016-2022) associations of blood p-tau181, glial fibrillary acidic protein (GFAP), neurofilament light (NfL), and amyloid-{beta}42/40 with memory, executive function, language, visuospatial ability, and global cognition using individual biomarker, principal components analysis-derived composite, and multibiomarker panel models. RESULTS: Cross-sectionally, NfL and GFAP showed the broadest associations. Prospectively, p-tau181 was independently associated with memory and global cognition, whereas GFAP was associated with executive function, memory, and global cognition. P-tau181 also showed relative memory-versus-executive s
- Tollecode
A local AI coding assistant to delegate tasks to AI agents
- FLUIX AI
AI agents for data center cooling and power
- Coinbase is laying off 14% of staff, citing AI. Read the letter from its CEO.
Coinbase is laying off 14% of staff, citing AI. Read the letter from its CEO. Business Insider
- Tiny teams, no 'pure managers,' and AI: Coinbase's memo has all the classic ingredients of a 2026 layoff letter
Tiny teams, no 'pure managers,' and AI: Coinbase's memo has all the classic ingredients of a 2026 layoff letter Business Insider
- A lawyer reveals how he snagged $9 million in 4 days for his AI-native law firm
A lawyer reveals how he snagged $9 million in 4 days for his AI-native law firm Business Insider
- U.S. Commerce Department Agency to Test Google, Microsoft, xAI Models
U.S. Commerce Department Agency to Test Google, Microsoft, xAI Models The Information
- Coinbase to Cut 14% Jobs, Citing Market Conditions and AI
Coinbase to Cut 14% Jobs, Citing Market Conditions and AI The Information
- Pennsylvania is suing Character.AI after a chatbot posed as a licensed psychiatrist
A state investigator posing as a patient found a Character.AI chatbot claiming to be licensed in Pennsylvania and providing a fake license number
- US and tech firms strike deal to review AI models for national security before public release
Microsoft, Google DeepMind and xAI products to be vetted for cybersecurity, biosecurity and chemical weapons risks The US government has struck deals with Google DeepMind, Microsoft and xAI to review early versions of their new AI models before they are released to the public. The Center for AI Standards and Innovation (CAISI), part of the US Department of Commerce, announced the agreements on Tuesday, saying the review process would be key to understanding the capabilities of new and powerful AI models as well as to protecting US national security. These collaborations will help the federal government “scale (its) work in the public interest at a critical moment”, the agency said in a press release. Continue reading...
- US to safety test new AI models from Google, Microsoft, xAI
New agreements between the companies and the Commerce department build on Biden-era pacts.
- Y Combinator alum Moritz raises $9m to build a law firm supercharged with AI
Y Combinator alum Moritz raises $9m to build a law firm supercharged with AI
- Coinbase CEO Brian Armstrong in layoff email to employees: I have watched engineers use AI tools
Coinbase is cutting about 14% of its global workforce, impacting around 700 jobs. This move is driven by the evolving role of artificial intelligence in operations and ongoing crypto market fluctuations. The company aims to streamline its structure and enhance efficiency. Restructuring is expected to conclude by the second quarter of 2026.
- White House briefed Anthropic, Google, and OpenAI on plans for a government AI review process
After a year of deregulation, the White House is now discussing an executive order that could subject new AI models to government review before they are released. The trigger is said to be Anthropic's "Mythos" model. The article White House briefed Anthropic, Google, and OpenAI on plans for a government AI review process appeared first on The Decoder .
- CAISI Signs Agreements Regarding Frontier AI National Security Testing With Google DeepMind, Microsoft and xAI
Expanded collaborations with leading U.S. AI labs provide for pre-deployment evaluations and other research.
- Coinbase slashes headcount, sees "fleets of agents" instead
"Non-technical teams are now shipping production code and many of our workflows are being automated"
- A chatbot told a state investigator it was a licensed psychiatrist. It gave a fake licence number. Pennsylvania just sued.
A state investigator in Pennsylvania created an account on Character.AI, opened a conversation with a chatbot called Emilie, and told it he was feeling depressed. Emilie responded that she was a psychiatrist, that she had attended Imperial College London’s medical school, that she was licensed to practise in Pennsylvania and the United Kingdom, and that […] This story continues at The Next Web
- Coinbase cuts 14pc of jobs to save costs and embrace AI
Last year, Coinbase Europe was fined nearly €21.5m for failing to monitor transactions. Read more: Coinbase cuts 14pc of jobs to save costs and embrace AI
- AIProofMe
Get your free AI-proof career score.
- Vyrill Agentic Video Commerce Platform
Agentic infrastructure powering video search & monetization.
- Stochastic KV Routing: Enabling Adaptive Depth-Wise Cache Sharing
Serving transformer language models with high throughput requires caching Key-Values (KVs) to avoid redundant computation during autoregressive generation. The memory footprint of KV caching is significant and heavily impacts serving costs. This work proposes to lessen these memory requirements. While recent work has largely addressed KV cache reduction via compression and eviction along the temporal axis, we argue that the depth dimension offers an orthogonal and robust avenue for optimization. Although prior research suggests that a full cache for every layer is redundant, implementing…
- Label-Efficient School Detection from Aerial Imagery via Weakly Supervised Pretraining and Fine-Tuning
Accurate school detection is essential for supporting education initiatives, including infrastructure planning and expanding internet connectivity to underserved areas. However, many regions around the world face challenges due to outdated, incomplete, or unavailable official records. Manual mapping...
- TabSurv: Adapting Modern Tabular Neural Networks to Survival Analysis
Survival analysis on tabular data is a well-studied problem. However, existing deep learning methods are often highly task-specific, which can limit the transfer of new approaches from other domains and introduce constraints that may affect performance. We propose TabSurv, an approach that adapts mo...
- Atomic Fact-Checking Increases Clinician Trust in Large Language Model Recommendations for Oncology Decision Support: A Randomized Controlled Trial
Question: Does atomic fact-checking, which decomposes AI treatment recommendations into individually verifiable claims linked to source guideline documents, increase clinician trust compared to traditional explainability approaches? Findings: In this randomized trial of 356 clinicians generating 7...
- Steer Like the LLM: Activation Steering that Mimics Prompting
Large language models can be steered at inference time through prompting or activation interventions, but activation steering methods often underperform compared to prompt-based approaches. We propose a framework that formulates prompt steering as a form of activation steering and investigates wheth...
- QKVShare: Quantized KV-Cache Handoff for Multi-Agent On-Device LLMs
Multi-agent LLM systems on edge devices need to hand off latent context efficiently, but the practical choices today are expensive re-prefill or full-precision KV transfer. We study QKVShare, a framework for quantized KV-cache handoff between agents that combines token-level mixed-precision allocati...
- Spatiotemporal Convolutions on EEG signal -- A Representation Learning Perspective on Efficient and Explainable EEG Classification with Convolutional Neural Nets
Classification of EEG signals using shallow Convolutional Neural Networks (CNNs) is a prevalent and successful approach across a variety of fields. Most of these models use independent one-dimensional (1D) convolutional layers along the spatial and temporal dimensions, which are concatenated without...
- EvoLM: Self-Evolving Language Models through Co-Evolved Discriminative Rubrics
Language models encode substantial evaluative knowledge from pretraining, yet current post-training methods rely on external supervision (human annotations, proprietary models, or scalar reward models) to produce reward signals. Each imposes a ceiling. Human judgment cannot supervise capabilities be...
- Correct Is Not Enough: Training Reasoning Planners with Executor-Grounded Rewards
Reinforcement learning with verifiable rewards has become a common way to improve explicit reasoning in large language models, but final-answer correctness alone does not reveal whether the reasoning trace is faithful, reliable, or useful to the model that consumes it. This outcome-only signal can r...
- MCJudgeBench: A Benchmark for Constraint-Level Judge Evaluation in Multi-Constraint Instruction Following
Multi-constraint instruction following requires verifying whether a response satisfies multiple individual requirements, yet LLM judges are often assessed only through overall-response judgments. We introduce MCJudgeBench, a benchmark for constraint-level judge evaluation in multi-constraint instruc...
- Mechanical Conscience: A Mathematical Framework for Dependability of Machine Intelligenc
Distributed collaborative intelligence (DCI), encompassing edge-to-edge architectures, federated learning, transfer learning, and swarm systems, creates environments in which emergent risk is structurally unavoidable: locally correct decisions by individual agents compose into globally unacceptable ...
- TRACE: A Metrologically-Grounded Engineering Framework for Trustworthy Agentic AI Systems in Operationally Critical Domains
We introduce TRACE, a cross-domain engineering framework for trustworthy agentic AI in operationally critical domains. TRACE combines a four-layer reference architecture with an explicit classical-ML vs. LLM-validator split (L2a/L2b), a stateful orchestration-and-escalation policy (L3), and bounded ...
- RoboAlign-R1: Distilled Multimodal Reward Alignment for Robot Video World Models
Existing robot video world models are typically trained with low-level objectives such as reconstruction and perceptual similarity, which are poorly aligned with the capabilities that matter most for robot decision making, including instruction following, manipulation success, and physical plausibil...
- ScrapMem: A Bio-inspired Framework for On-device Personalized Agent Memory via Optical Forgetting
Long-term personalized memory for LLM agents is challenging on resource-limited edge devices due to high storage costs and multimodal complexity. To address this, we propose ScrapMem, a framework that integrates multimodal data into "Scrapbook Page." ScrapMem introduces Optical Forgetting, an optica...
- Mona
Automating business capital access with AI agents
- AI Advocate: Educational Path to Transform Squads to the Future
This paper analyzes the strategic education process aimed at transitioning traditional software development squads into hybrid structures centered on collaborative work between humans and Artificial Intelligence (AI). In a context where human-AI collaboration can significantly increase productivity,...
- Say the Mission, Execute the Swarm: Agent-Enhanced LLM Reasoning in the Web-of-Drones
Large Language Models (LLMs) are increasingly explored as high-level reasoning engines for cyber-physical systems, yet their application to real-time UAV swarm management remains challenging due to heterogeneous interfaces, limited grounding, and the need for long-running closed-loop execution. This...
- OracleProto: A Reproducible Framework for Benchmarking LLM Native Forecasting via Knowledge Cutoff and Temporal Masking
Large language models are moving from static text generators toward real-world decision-support systems, where forecasting is a composite capability that links information gathering, evidence integration, situational judgment, and action-oriented decision making. This capability is in broad demand a...
- Rethinking the Rank Threshold for LoRA Fine-Tuning
A recent landscape analysis of LoRA fine-tuning in the neural tangent kernel regime establishes a sufficient condition $r(r+1)/2 > KN$ on the LoRA rank $r$ for the absence of spurious local minima under squared-error loss, prescribing $r \geq 12$ on canonical few-shot RoBERTa setups. The condition i...
- Segmenting Human-LLM Co-authored Text via Change Point Detection
The rise of large language models (LLMs) has created an urgent need to distinguish between human-written and LLM-generated text to ensure authenticity and societal trust. Existing detectors typically provide a binary classification for an entire passage; however, this is insufficient for human--LLM ...
- SAM-NER: Semantic Archetype Mediation for Zero-Shot Named Entity Recognition
Zero-shot Named Entity Recognition (ZS-NER) remains brittle under domain and schema shifts, where unseen label definitions often misalign with a large language model's (LLM's) intrinsic semantic organization. As a result, directly mapping entity mentions to fine-grained target labels can induce syst...
- SERE: Structural Example Retrieval for Enhancing LLMs in Event Causality Identification
Event Causality Identification (ECI) requires models to determine whether a given pair of events in a context exhibits a causal relationship. While Large Language Models (LLMs) have demonstrated strong performance across various NLP tasks, their effectiveness in ECI remains limited due to biases in ...
- Tailored Prompts, Targeted Protection: Vulnerability-Specific LLM Analysis for Smart Contracts
Smart contracts on blockchains are prone to diverse security vulnerabilities that can lead to significant financial losses due to their immutable nature. Existing detection approaches often lack flexibility across vulnerability types and rely heavily on manually crafted expert rules. In this paper, ...
- Graph Neural Network based Hierarchy-Aware Embeddings of Knowledge Graphs: Applications to Yeast Phenotype Prediction
We present a method for finding hierarchy-aware embeddings of knowledge graphs (KGs) using graph neural networks (GNNs) enriched with a semantic loss derived from underlying ontologies. This method yields embeddings that better reflect domain knowledge. To demonstrate their utility, we predict and i...