Build and evaluate a disease forecast from surveillance data
See the requirement-level assessment and its sources.
SCIENCE UNDER AUTOMATION
AI in Population health: capabilities and evidence
Population health connects observations and forecasts to interventions whose effects must be measured in people.
Autonomy · up to 4.4 / 10
Model interval: 2.1–4.4. Assessed 21 October 2026.
Two equally weighted workflow families: a bounded operational task and an investigation or deployment beyond that task. Inherited subfields use this parent scenario.
Calculation and work mix
Each requirement uses the operational scale from 0 (independent execution unestablished) to 4 (reliable independent execution in the stated setting). A workflow takes its lowest requirement level. The two workflows receive equal weight; the headline is 1 + 9 × weighted level / 4.
These levels are source-linked model assignments, not measurements reported by the papers. “Up to” is the upper model bound, not a probability or a forecast date.
Build and evaluate a disease forecast from surveillance data: 1–2 / 4.
Design, deliver and evaluate a population intervention: 0–1 / 4.
Upper bound with 75:25 workflow weights: 4.9. With 25:75 weights: 3.8.
What has AI demonstrated?
ERA generated forecasting software during the 2025–26 US respiratory season. A selected ensemble ranked first overall in the authors’ report of official post-season evaluation.
- Test
- Primary preprint; prospective real-time evaluation
- Comparison
- Comparators vary by study; see the linked sources.
- System and version
- Prospective multi-pathogen forecasting, v2 · 2026-10-14
- Evidence checked
- 2026-10-21
Methodological approaches were usually supplied. Surveillance data and model selection remain part of the configured forecasting system; public-health interventions were not automated.
What does the field require?
Build and evaluate a disease forecast from surveillance data
- Generate and revise forecasting software
Autonomy level: 2–3 / 4.
ERA builds executable models, usually from supplied methodological directions.
Prospective multi-pathogen forecasting, v2 - Evaluate predictions prospectively
Autonomy level: 2–3 / 4.
The prospective season evaluation tests forecasts on outcomes unavailable when predictions were made.
Prospective multi-pathogen forecasting, v2 · Real-time respiratory outbreak early warning - Maintain a decision-ready forecasting pipeline
Autonomy level: 1–2 / 4.
Data feeds, operational integration and decision-makers remain part of the surrounding infrastructure.
Real-time respiratory outbreak early warning
Design, deliver and evaluate a population intervention
- Define a causal comparison and recruit participants
Autonomy level: 0–1 / 4.
The field trial used a researcher-designed randomization and enrolled patients through clinical facilities.
Generative AI decision support in primary care - Implement the intervention across settings
Autonomy level: 0–1 / 4.
People and institutions delivered the intervention; forecasting systems supply predictions.
Generative AI decision support in primary care · Real-time respiratory outbreak early warning - Interpret outcomes against bias and competing explanations
Autonomy level: 1–2 / 4.
Statistical and predictive tools assist analysis, but forecasts alone do not identify intervention effects.
Generative AI decision support in primary care · Prospective multi-pathogen forecasting, v2
Research tasks
Design, deliver and evaluate a population intervention
See the requirement-level assessment and its sources.
Can AI replace the whole workflow?
The results above measure specific tasks. The field assessment also requires completion of the connected workflow, including the requirements listed here. A field-wide replacement percentage has not been established.
Read the assessment method