The applications of artificial intelligence seem to be almost endless, especially in this historical period strongly characterized by technological innovation and the growing interest in automation. Amazon Web Services’ latest product, HealthScribe, is a new generative artificial intelligence system designed to help doctors compile clinical documentation that promises to save professionals up to 6 hours a day. Yes…but how?
Amazon Web Services (AWS) is the cloud system created by the Seattle giant in 2006. Since then, AWS has continuously experimented and included in its offer the most diverse services aimed at reducing the workload for professionals in various sectors.
From data processing to storage, from database management to augmented reality. As anticipated, the latest system developed by Amazon (or rather, AWS) is intended for medical staff with the aim of reducing the daily effort required for the compilation of patient medical records and the related data entry in electronic medical records.
Benefits: more time available to patients and less risk of burnout. Generative AI has found great application in various professional fields and promises great results in the health sector as well.
Amazon AWS, HealthScribe, and clinical documentation
Big promises and lots of expectations, but how does it work?
This new system combines speech recognition with generative artificial intelligence to automatically create preliminary clinical documentation from doctor-patient conversations.
Speech recognition allows the professional to free themselves from the burden of transcribing the conversations, automatically noting what has been said by both parties during medical examinations.
This technology is widely used by users around the world in various fields of application.
Over the years, numerous technology companies have developed systems of this type, designed to best meet the needs of users in different sectors of use. These include AWS, which in 2019 launched its own Transcribe Medical, able to report and track the information provided to the doctor by individual patients.
The next step offered by HealthScribe is enabled using a generative artificial intelligence system capable of processing transcripts and independently writing electronic medical records, extracting only the relevant information, and identifying the key terms for diagnosis.
However, there is no shortage of critical issues highlighted by various stakeholders, starting with the lack of data relating to the accuracy of the service. The concerns also include the possible biases involved in this type of system and the dangers resulting from a lack of supervision of what has been transcribed and processed.
For now, we can only await further details.
Image by Laura James from Pexels