Editing AI for Epidemiology (section)

== <span style="color: #FFFFFF;">Analyzing</span> ==
{| class="wikitable"
|+ Epidemiology AI Performance vs. Traditional Methods
! Application !! AI Advantage !! Key Limitation !! Reliability
|-
| Digital surveillance (flu) || 1-2 week lead time || Fragile correlations || Moderate
|-
| Wastewater surveillance || 7-10 day early warning || Catchment area complexity || High
|-
| Epidemic forecasting (short-term) || Comparable to statistical || Non-stationary dynamics || Moderate
|-
| Genomic variant tracking || Automated, fast || Sequencing bias || High
|-
| Individual risk models (EHR) || Personalized prediction || Validation quality varies || Variable
|}

'''Failure modes''': Distribution shift during pandemic phases (model trained pre-Omicron fails post-Omicron). Data quality collapses during surges (under-reporting, delayed reporting). Algorithmic amplification of disparities — risk models trained on historically under-served populations perform worse precisely where intervention is most needed. Overconfident point predictions without uncertainty quantification leading to poor public health decisions.
</div>

<div style="background-color: #483D8B; color: #FFFFFF; padding: 20px; border-radius: 8px; margin-bottom: 15px;">