Research led by Mass General Brigham reveals that the actual prevalence of long COVID may be double the current estimates, affecting about 16% of COVID-19 patients in the U.S., which translates to over 18 million individuals. This study analyzed medical records of nearly 460,000 patients across 58 hospitals using a novel AI algorithm designed to identify long COVID through longitudinal health data. Current diagnostic codes, including the ICD code U09.9 for post-COVID conditions, capture fewer than 7% of long COVID cases, resulting in many individuals remaining undiagnosed. The study found significant regional variations in long COVID manifestations, with some areas showing higher rates of conditions like prediabetes. Importantly, the prevalence of long COVID continues to increase, suggesting that the virus is still contributing to new chronic health conditions. The research underscores the potential of clinical AI in enhancing public health surveillance and improving identification of complex post-viral conditions.