Azernews.Az

By Alimat Aliyeva

According to a joint study conducted by scientists from Harvard and Stanford Universities, artificial intelligence (AI) can make more accurate diagnoses than humans, particularly under stressful conditions, Azernews reports.

"The accuracy of diagnoses made by artificial intelligence when admitting patients to the intensive care unit reached 79.7 percent, compared to 75.9 percent for leading specialists. AI proved especially effective in high-pressure situations where there is very little time for a detailed review of the patient’s entire medical history," the study reports.

Misdiagnoses are responsible for approximately 10 percent of patient deaths worldwide. In the United States, misdiagnosis ranks among the leading causes of mortality. In light of this, researchers have attempted to integrate AI as a diagnostic assistant to support medical professionals. However, this approach has yet to significantly improve diagnostic accuracy, largely because doctors often distrust or hesitate to rely on AI-generated conclusions.

Despite its proven diagnostic potential, one of the biggest hurdles for AI in healthcare remains human acceptance and trust. Bridging this gap may require not only technological advances but also improved collaboration frameworks, transparency in AI decision-making, and enhanced education for medical practitioners about AI capabilities and limitations. As AI continues to evolve, its role might shift from consultant to co-decision maker, fundamentally transforming how healthcare is delivered.