https://theimpactofai.net/story/breast-ultrasound-ai-clinical-context-256-patients-2026 The Impact of AIen 2026-10-19T04:04:04.000Z Can clinical context improve breast-ultrasound AI? A 256-patient test is not decisive https://theimpactofai.net/story/us-agencies-ai-agent-incident-response-practices-2026 The Impact of AIen 2026-10-19T03:01:22.000Z US agencies set ten AI-agent incident practices—what changes for operators? https://theimpactofai.net/story/anthropic-public-benefit-fund-five-percent-shares-2026 The Impact of AIen 2026-10-19T01:04:39.000Z Anthropic pledged 5% of its shares to public benefit—what is guaranteed? https://theimpactofai.net/story/ai-lipopeptide-antibiotic-mouse-wounds-2026 The Impact of AIen 2026-10-18T13:03:34.000Z Did AI design an antibiotic that worked in mice? https://theimpactofai.net/story/wifi-localisation-fewer-router-signals-one-building-2026 The Impact of AIen 2026-10-18T12:01:57.000Z A Wi-Fi locator discarded up to 95% of router signals—inside one building https://theimpactofai.net/story/anthropic-public-staff-ai-expectations-gap-2026 The Impact of AIen 2026-10-18T09:57:36.000Z Anthropic staff expect far more AI automation than US workers—what does that prove? https://theimpactofai.net/story/reflection-beam-open-weights-safety-evidence-2026 The Impact of AIen 2026-10-18T09:02:04.000Z Reflection released Beam's weights—what can outsiders verify? https://theimpactofai.net/story/deepseek-v41-flash-1017-company-benchmarks-2026 The Impact of AIen 2026-10-18T07:04:54.000Z DeepSeek's open-weight update won some agent tests—who ran them? https://theimpactofai.net/story/imf-asia-ai-job-postings-relative-decline-2026 The Impact of AIen 2026-10-18T05:25:30.000Z AI-exposed job adverts fell in Asia—but 10% is a relative estimate https://theimpactofai.net/story/ai-organ-scans-hcc-immunotherapy-association-2026 The Impact of AIen 2026-10-18T02:00:58.000Z Routine scans revealed liver-cancer survival signals—but not a treatment rule https://theimpactofai.net/story/alzheimer-icap-blood-test-exploratory-2026 The Impact of AIen 2026-10-18T02:00:58.000Z An Alzheimer’s blood test found a signal—the blind test had 23 people https://theimpactofai.net/story/federated-causal-healthcare-review-2026 The Impact of AIen 2026-10-18T02:00:58.000Z Keeping patient data local does not make an AI health claim causal https://theimpactofai.net/story/factory-edge-ai-energy-one-line-2026 The Impact of AIen 2026-10-18T01:03:23.000Z A factory reported 26% lower energy use with AI—but only on one line https://theimpactofai.net/story/malware-classifier-benchmark-not-live-attacks-2026 The Impact of AIen 2026-10-18T01:03:23.000Z A malware classifier hit 100% on a benchmark—not on live attacks https://theimpactofai.net/story/cancer-histology-hierarchical-model-second-dataset-gap-2026 The Impact of AIen 2026-10-18T00:15:00.000Z A cancer-image model scored 98.5%—then fell to 82.0% on a second dataset https://theimpactofai.net/story/solar-forecast-zero-output-postprocessing-turkiye-2026 The Impact of AIen 2026-10-18T00:15:00.000Z A solar forecast fix cut percentage error—mostly by learning when output is zero https://theimpactofai.net/story/llm-depression-diary-signal-not-diagnosis-2026 The Impact of AIen 2026-10-17T23:01:00.000Z LLMs found depression signals in one-minute diaries—but not a diagnosis https://theimpactofai.net/story/lung-nodule-model-cross-cohort-calibration-2026 The Impact of AIen 2026-10-17T23:01:00.000Z A lung-nodule model travelled better—but its labels were not pathology https://theimpactofai.net/story/llm-spectra-features-random-forest-better-2026 The Impact of AIen 2026-10-17T21:59:24.000Z LLMs read compact spectra better—but a random forest still won https://theimpactofai.net/story/stock-model-calendar-leakage-shap-explanation-2026 The Impact of AIen 2026-10-17T21:59:24.000Z A stock model scored 63.84%—a calendar rule did better https://theimpactofai.net/story/liver-cancer-ai-risk-korea-low-base-rate-2026 The Impact of AIen 2026-10-17T21:03:56.000Z A liver-cancer model ranked risk well—but the cancer rate was 0.062% https://theimpactofai.net/story/oncology-ai-consultation-prep-small-workflow-study-2026 The Impact of AIen 2026-10-17T21:03:56.000Z AI summaries cut perceived oncology workload—but 12 of 33 were incomplete https://theimpactofai.net/story/dashguard-one-dataset-crash-anticipation-2026 The Impact of AIen 2026-10-17T18:00:12.000Z A dashcam model anticipated crashes on one dataset—what still needs proving? https://theimpactofai.net/story/hcv-ai-genotyping-standard-tool-comparison-2026 The Impact of AIen 2026-10-17T18:00:12.000Z Can a lighter AI genotype hepatitis C as well as the standard tool? https://theimpactofai.net/story/renewable-grid-anomaly-accuracy-zero-recall-2026 The Impact of AIen 2026-10-17T18:00:12.000Z A grid anomaly model scored 95% accuracy—but caught none of the faults https://theimpactofai.net/story/anthropic-metr-embedded-evaluators-access-2026 The Impact of AIen 2026-10-17T15:59:43.000Z Anthropic gave six METR staff internal access—what can they publish? https://theimpactofai.net/story/openai-rosalind-academic-plan-us-labs-2026 The Impact of AIen 2026-10-17T15:59:43.000Z OpenAI is offering GPT‑Rosalind to US labs—what is actually free? https://theimpactofai.net/story/ai-training-storage-agent-checkpoint-latency-2026 The Impact of AIen 2026-10-17T14:58:40.000Z A storage agent cut AI checkpoint delays—but only on one test platform https://theimpactofai.net/story/ai-writing-detectors-false-positive-review-2026 The Impact of AIen 2026-10-17T12:04:00.000Z An AI-writing detector should not decide whether a student cheated https://theimpactofai.net/story/large-medical-ai-bias-review-2026 The Impact of AIen 2026-10-17T09:56:08.000Z Medical AI bias research is growing—but it rarely reaches real clinics https://theimpactofai.net/story/pasture-ai-new-region-generalisation-2026 The Impact of AIen 2026-10-17T09:56:08.000Z Can a pasture AI travel to a new region? This one often could not