Research library
Medicine and clinical care
View the four-date completion graphs and reconstruction notes.
Promise and evidence
AI could improve diagnosis, treatment selection, workflow efficiency, and access to care. The MASAI randomized mammography study provides an example of prospective evaluation of AI-supported clinical practice. Diagnostic accuracy, clinician workload, and patient outcomes need separate measurement.
Apply the shared measurement protocol. Targets are proposals for evaluation, not clinical guidance.
Dated baseline coverage
All linked measurable-outcome files below contain an evidence baseline reviewed November 7, 2026, with source dates, findings, limitations and follow-up measurements. These distinguish measured results from proxies and unknown values; they do not assert that proposed milestones are achieved. See the baseline coverage notes.
Progress estimates
Judgment-based estimates as of November 7, 2026, using the same 0–100% scale and five-part rubric as mathematics. 0% anchors the period before ChatGPT's public release in 2022; 100% requires AI to manage this field's continuing advancement without human participation. Confidence is low. These are evidence-based judgments, not empirical completion measurements, forecasts or percentages of jobs replaced.
| Measurable outcome | Estimated progress | Execution / breadth / agenda / validation / improvement (each out of 20) |
|---|---|---|
| Care delivery | 25% | 10 / 5 / 5 / 5 / 0 |
| Cost effectiveness | 15% | 5 / 5 / 0 / 5 / 0 |
| Diagnostic performance | 35% | 15 / 5 / 5 / 10 / 0 |
| Equity and transfer | 15% | 5 / 5 / 0 / 5 / 0 |
| Harm and unnecessary care | 20% | 10 / 5 / 0 / 5 / 0 |
| Patient benefit | 15% | 5 / 5 / 0 / 5 / 0 |
Each linked outcome explains its score, cites its dated evidence baseline and identifies the next measurement. Zero indicates no qualifying post-anchor progress established by the reviewed evidence for that outcome, not proof of no relevant capability. Do not average these overlapping indicators into a field-wide percentage or interpret their differences as a precise ranking of fields. All outcomes reaching 100% would still require a separate field-wide demonstration of sustained autonomous advancement.
Measurable features
| Feature | Measurement and acceptance rule |
|---|---|
| Diagnostic performance | Sensitivity, specificity, positive predictive value, calibration, and missed diagnoses against an independently determined reference standard. Report prevalence and follow-up. |
| Patient benefit | Mortality, morbidity, functional outcomes, or quality of life at a specified follow-up, depending on the clinical problem. |
| Harm and unnecessary care | Adverse events, false-positive procedures, missed urgent cases, and inappropriate treatments per eligible patient. |
| Care delivery | Time to appropriate treatment, waiting time, completed visits, and clinician minutes per patient. Include time spent correcting AI output. |
| Equity and transfer | Outcome and error differences across relevant demographic groups, institutions, devices, and disease severities. |
| Cost effectiveness | Full cost per appropriately managed patient and, where valid, cost per quality-adjusted life-year gained. State assumptions. |
Evidence requirements
Evaluate the complete clinical workflow prospectively, preferably with randomized allocation where feasible. Compare with current care, not an artificially weak clinician baseline. Separate AI alone from clinician-plus-AI. Prespecify clinically meaningful differences and non-inferiority margins, use adequate power, and conduct follow-up sufficient to detect delayed harms. A higher detection rate alone does not establish longer survival.
Proposed milestones
Validate a frozen system prospectively at three independent sites, maintaining prespecified diagnostic and subgroup performance limits.
In a powered comparative trial, reduce clinician work by at least 20% while meeting a preregistered non-inferiority margin for clinical outcomes and harms.
Demonstrate a clinically meaningful improvement in the primary patient outcome, with uncertainty and absolute event rates reported.
Sustain the benefit over 12 months of deployment with monitoring for drift, errors, and total costs. This measures realized human benefit; claims about singularity feedback require additional evidence that research and innovation themselves accelerate.