The article discusses how a Nigerian doctor and AI specialist named Tobi Olatunji is using artificial intelligence (AI) to help African hospitals digitize medical records faster and save time for doctors. In Africa, there is a shortage of healthcare personnel, leading to doctors being overworked and patients facing long wait times. Olatunji co-founded a startup called Intron Health, which initially offered electronic health record (EHR) software but later developed a proprietary speech recognition platform.

Traditional EHR systems are not widely used in African healthcare due to cost and usability issues. Doctors find them cumbersome and time-consuming, leading to resistance in adoption. Intron Health addressed this challenge by developing a speech-to-text software that understands African accents, enabling doctors to dictate their notes instead of typing them. They collected a large dataset of African-accented voice samples and built automatic speech recognition software that outperforms popular speech recognition models like Google Assistant or Siri. Several hospitals in Nigeria, Kenya, Ghana, and South Africa are now using Intron's speech-to-text software.

Intron Health has partnered with GPU chipmaker Nvidia and has managed to build and maintain the infrastructure required for running an AI company. They are also running a developer challenge called Afri-Speech-200 to advance diversity in AI and encourage programmers to build better models using a portion of Intron's training dataset.

While Intron's speech recognition technology shows promise in improving efficiency for African healthcare workers, it also highlights the need to ensure that AI systems consider the unique contexts and accents outside of developed countries where training data is abundant.

Read more on TechCabal