AI-Designed Phages Reach a Patient
A company reports compassionate-use administration after its phage cocktail cleared the patient's bacterial isolate in vitro; the clinical outcome is undisclosed.
Read this GVC →Grains of Vital Cognizance by Georgi V. Chaltikyan
GVCs
One consequential development. One evidence-led thesis. One question that invites a better conversation.
Latest entries
Human Oversight of Clinical AI: The Error-Interception Test
A review finds that supervising AI is protective only when clinicians can recognize and correct its errors.
Read this GVC →
Personalized mRNA Cancer Therapy: The Phase III Readout
Detailed results show lower melanoma recurrence risk with an individualized therapy plus pembrolizumab; overall-survival evidence is still immature.
Read this GVC →
1,946 AI Scientists, One Biological Experiment
GPT-Rosalind agents analyzed millions of mouse brain cells in 32 minutes, producing a reasoning atlas of testable hypotheses.
Read this GVC →
MAI-DxO: From Benchmark Toward Clinical Testing
The 2025 diagnostic AI now has a peer-reviewed paper, broader evaluation, released research tools and a planned clinical study.
Read this GVC →
Longevity Needs CUDOS, Not DECAY
A BMJ opinion challenges longevity hype; rigorous science can protect the hope of healthier ageing.
Read this GVC →
AI Biology’s Dual-Use Dilemma
A BMJ editorial asks how biosecurity can keep pace with AI-enabled biology while preserving its medical promise.
Read this GVC →
Training Doctors for the AI Era
A JAMA Viewpoint warns of “never-skilling” and argues for physicians who can reason with AI—and without it.
Read this GVC →
The $1.8B Biology Data Push
Biohub and partners aim to build standardized experimental data for predictive models of cells—and make biological measurement the priority.
Read this GVC →
When an AI Agent Crossed a Government Boundary
An agent seeking health-spending data gained unauthorised access to an Australian statistics portal—showing why action boundaries matter.
Read this GVC →
Claude Gets a Pipette: AI Moves into the Wet Lab
Anthropic’s lab and verified-access programme bring frontier AI closer to the experimental cycle—while real biology remains the test.
Read this GVC →
Can Medical AI Leapfrog Health-System Infrastructure?
Anthropic and OpenEvidence plan free clinical AI in roughly 100 countries—but local relevance will be as important as access.
Read this GVC →
Who Evaluates Digital Health? The Developer Effect
A review of 229 trials finds developer involvement is common and associated with more statistically significant results.
Read this GVC →
Responsible AI for Health Moves from Principles to Implementation
GI-AI4H is turning high-level principles into standards, benchmarking and institutional capacity for real health systems.
Read this GVC →
Clinical AI Learns When to Act—and When to Defer
On-premise AI combines institutional control with decision-time reliability to support selective autonomy and human escalation.
Read this GVC →
From Static Paper to Executable Research Agent
Paper2Agent turns publications, code, data and workflows into interactive agents that can execute and reuse scientific methods.
Read this GVC →
Scientific AI Enters Real R&D Workflows
GPT-Rosalind moves beyond research preview toward governed, connected scientific work inside R&D organizations.
Read this GVC →
From Organ Model to Whole-Person Digital Twin
A new international consortium aims to turn medical digital twins into longitudinal infrastructure for precision health.
Read this GVC →
Health Age Goes Mainstream
Continuous sensing, intelligent interpretation and longevity framing converge in a mass-market consumer ecosystem.
Read this GVC →
Autonomous Clinical AI Enters a Federally Funded Regulatory Pathway
ADVOCATE puts funding, safety supervision, health-system validation and an FDA pathway behind clinical agentic AI.
Read this GVC →
Every Single-Letter Genome Change Is Now Searchable
AlphaGenome Atlas makes predictions for roughly nine billion possible variants searchable at genome-wide scale.
Read this GVC →
ChatGPT Moves Closer to the Patient Chart
Authorized patient context, evidence and clinical workflow move toward an intelligent clinical workspace.
Read this GVC →
Longevity Clocks Enter Clinical Trials
Six proteomic ageing clocks bring exploratory biological-age signals into a phase 2a intervention study.
Read this GVC →
A Smartphone Window on Brain Disease
Remote cognitive assessments show promising links with brain atrophy in early Alzheimer’s disease.
Read this GVC →
Governance Defines AI Readiness
Responsible progress in AI for health begins with institutions capable of governing it safely.
Read this GVC →
AI Enters the Live Operative Field
Real-time AI interpreted the live endoscopic field while the human surgical team remained in control.
Read this GVC →
A Licence for AI Doctors?
Reported industry discussions about a national licence reveal where autonomous care is pushing regulation.
Read this GVC →
Measured Does Not Always Mean Inferred Correctly
A precise sensor reading does not automatically produce a precise health interpretation.
Read this GVC →
Living Neurons Have Entered the Server Rack
Wetware has entered a live computing environment—but useful superiority over conventional hardware remains unproven.
Read this GVC →
Can AI Become a Virtual Laboratory for Biology?
The promise is not replacing experiments, but deciding which experiments deserve to be done.
Read this GVC →
Can We Test Longevity Interventions Without Waiting Decades?
Responsive epigenetic clocks could accelerate longevity trials—but a younger clock is not proof of rejuvenation.
Read this GVC →
Consumer AI Is Becoming Healthcare’s Front Door
When AI shapes routing, records and transactions, it becomes healthcare infrastructure.
Read this GVC →
A Robot Can Draw Blood. What Comes Next?
The first standalone robotic blood-draw device brings supervised procedural autonomy into routine outpatient care.
Read this GVC →
Has Biomedical Publishing Crossed the AI Point of No Return?
The important question may soon be not whether AI was used in biomedical publishing, but how.
Read this GVC →
Personalized mRNA Cancer Vaccines Reach a Phase 3 Milestone
The first positive late-stage trial for a personalized mRNA cancer vaccine may mark an important precision-oncology milestone.
Read this GVC →
Predict Less. Adapt More.
Healthcare needs adaptive planning, interoperable data and governance that can evolve alongside AI capability.
Read this GVC →
Three Signals in 48 Hours: Is Medicine Entering the Age of the AI Clinical Actor?
Three developments are shifting the debate from AI as a tool toward AI as a potentially regulated clinical actor.
Read this GVC →
AI Is Learning to See the Patient. What Remains Uniquely Human?
In a simulated video OSCE, AMIE outscored primary-care physicians on clinical rubrics and perception—but it is not ready for clinical deployment.
Read this GVC →
Social Experience May Enter the Body Through the Biological Clock
Positive social experiences and childhood adversity were associated with mortality and epigenetic ageing—but the observational evidence does not establish causality.
Read this GVC →
AI Cannot Be an Author. But Can the Human Claim Its Discovery?
JAMA keeps authorship accountability human; AI-generated mathematical results challenge whether human-only credit can still describe who made the discovery.
Read this GVC →
Why a Catastrophe Caused by Artificial Intelligence May Be Inevitable
Dual-use capability races can reduce danger but cannot eliminate residual risk; resilience must prepare for consequential failure.
Read this GVC →
The Genome Remembers Reproductive Timing
Parental age and individual reproductive procedures were associated with distinct de novo mutation patterns, but population-level genomic associations must not become individual blame.
Read this GVC →
The Benchmark Ends Where Clinical AI Deployment Begins
ChatEHR’s deployment shows why clinical AI must be monitored across real workflows, unanticipated uses, human responses, and patient impact—not only pre-deployment benchmarks.
Read this GVC →
AI Found a Way into an “Undruggable” Cancer Target—but Biology Had the Final Say
AI narrowed nearly 40,000 compounds to GIPCi, but biochemical, cellular, and animal validation—not prediction alone—made it a credible preclinical candidate.
Read this GVC →
Health Apps Work. The Evidence Still Wobbles.
App-based interventions improve multiple NCD outcomes, but methodological weakness means a positive signal is not the same as trustworthy evidence.
Read this GVC →
Health AI Needs Its Own “Safe Enough” Framework
Health AI needs lifecycle safety evaluation against measurable real-world care, not a meaningless benchmark of perfection.
Read this GVC →
Can Better DNA Repair Extend Healthspan?
Better genome maintenance may support healthier ageing—but promising does not mean proven.
Read this GVC →
A Risk Score Is a Snapshot. Health Is a Trajectory: 10P-Health in Action
Longitudinal modelling puts time back into personalised health—but prediction is not prevention.
Read this GVC →
Govern Health AI by the Decisions It Can Shape—not Only by the Benchmarks It Can Pass
Capability and authority are not the same: govern the decisions a deployed system can shape.
Read this GVC →
We’re Done Waiting: Interoperability Must Move from Policy to Execution
Interoperability’s bottleneck is no longer standards alone—it is coordinated execution.
Read this GVC →
Clinical AI Can Cause Harm—and Still Improve Physician Performance
The next safety frontier is how we design—and train—the human–AI team.
Read this GVC →
How Can We Make Health AI Safer?
Frequency matters: measure, escalate and compare preventable harm.
Read this GVC →
Who Broke the Healthcare System?
The healthcare system is shaped by institutions and leaders who set its rules—and must accept responsibility for reform.
Read this GVC →
Popular Peptide, Divided Regulatory Vote
Consumer demand may influence regulatory debate, but popularity cannot establish clinical efficacy or safety.
Read this GVC →
Humanoid Robots Have Entered the Operating Room
General-purpose humanoids have demonstrated preclinical surgical feasibility, but teleoperation is not autonomy.
Read this GVC →
194 Years? A Biological Model—Not a Longevity Promise
A theoretical ceiling can guide science—but it is not evidence that humans can reach it.
Read this GVC →
AI Is Becoming a Commercial Determinant of Health
AI is not only changing healthcare. It is changing how commercial power shapes health.
Read this GVC →
21% Fewer Deaths — and Still 1.16 Million Lost
Road-safety progress is real. The toll remains intolerable.
Read this GVC →
AI4H Sovereignty
Medical AI should be contextualised, validated and governed within the health system where it will be used.
Read this GVC →
Can Medical AI Forget?
Deleting patient data may not remove its influence from a trained medical AI model.
Read this GVC →
From Device to Care
Connected health is moving from monitoring toward clinical care and outcome-aligned payment.
Read this GVC →
No Automation Without Representation
Frontline clinicians need a real voice in how clinical AI is used.
Read this GVC →
One Trillion Minutes
One wearable foundation model transferred across 35 health-prediction tasks.
Read this GVC →
Younger—But Biologically Older
Later-born generations show greater biological ageing and a higher observed risk of early-onset cancer.
Read this GVC →
Evidence First
Traditional medicine must meet the same standards of efficacy and safety as every other intervention.
Read this GVC →
Protect the Craft
AI training on surgical expertise requires consent, transparency, ownership, and control.
Read this GVC →
Perverse Incentives
Health systems need incentives that reward prevention, outcomes, and long-term health.
Read this GVC →
Retraining
Physicians will increasingly need to act as scientific interpreters rather than protocol-followers.
Read this GVC →
A Blood Signature for Lung Cancer Risk
A 14-protein panel may identify elevated lung-cancer risk more than five years before diagnosis.
Read this GVC →
The $1.2 Quadrillion Opportunity
The economic value of extended healthspan could be enormous—but access will determine who benefits.
Read this GVC →
Who Drives Healthcare Costs?
A US survey finds that 81% of respondents identify corporate health insurers as the greatest influence.
Read this GVC →
RISE
Equitable AI for Health requires capacity, inclusive research, accessible infrastructure, and secure data practices.
Read this GVC →
AI Arms Race
When providers use AI to code and payers use AI to deny, the system may automate mistrust faster than care.
Read this GVC →
UK Biobank Data Leak
The exposure of participants’ medical details is a warning shot for health-data governance.
Read this GVC →
Financial Toxicity of Cancer
Cancer can carry severe financial consequences, including markedly higher bankruptcy risk and lower credit scores.
Read this GVC →
77 Cents for Biomedical Data
The gap between AI investment and sustained biomedical-data investment remains extraordinary.
Read this GVC →