Artificial intelligence could spur a revolution in medical technology development, but inaccuracies in electronic medical records could hinder progress in the field.
AI has the potential to harness big data sources—including electronic medical records as well as molecular data—to detect patterns that could improve diagnoses, prevention strategies, and treatment decisions.
However, electronic medical records present obstacles for researchers interested in using AI because of the lack of standards in the data quality.
“The bane of the existence of people in AI is dirty data,” retired Microsoft executive Craig Mundie said July 23.
For example, the practice of upcoding, or ...